The work-reflection-learning cycle - Department of Computer and ...
The work-reflection-learning cycle - Department of Computer and ...
The work-reflection-learning cycle - Department of Computer and ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Birgit Rognebakke Krogstie<br />
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong><br />
<strong>cycle</strong> in s<strong>of</strong>tware engineering<br />
student projects: Use <strong>of</strong><br />
collaboration tools<br />
<strong>The</strong>sis for the degree philosophiae doctor<br />
Trondheim, June 2010<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
Faculty <strong>of</strong> Information Technology, Mathematics <strong>and</strong><br />
Electrical Engineering<br />
<strong>Department</strong> <strong>of</strong> <strong>Computer</strong> <strong>and</strong> Information Science
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
NTNU<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
<strong>The</strong>sis for the degree doctor philosophiae<br />
Faculty <strong>of</strong> Information Technology, Mathematics <strong>and</strong> Electrical Engineering<br />
<strong>Department</strong> <strong>of</strong> <strong>Computer</strong> <strong>and</strong> Information Science<br />
© 2010 Birgit Rognebakke Krogstie<br />
ISBN 978-82-471-2025-5 (printed version)<br />
ISBN 978-82-471-2026-2 (electronic version)<br />
ISSN 1503-8181<br />
Doctoral thesis at NTNU, 2010:35<br />
Printed in Norway by NTNU-trykk, Trondheim<br />
ii
Figure 1: Most used terms in the thesis (the research papers excluded); diagram<br />
generated in Wordle (http://www.wordle.net)<br />
iii
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
iv
Abstract<br />
In project based <strong>learning</strong>, students learn from h<strong>and</strong>s-on experience with the challenges<br />
<strong>and</strong> complexities <strong>of</strong> real <strong>work</strong>. To achieve this <strong>learning</strong>, <strong>reflection</strong> is essential. In project<br />
<strong>work</strong> in industry, retrospective <strong>reflection</strong> on the project process is established practice,<br />
e.g. in project debriefings. Student projects <strong>of</strong>ten include retrospective <strong>reflection</strong> as a<br />
m<strong>and</strong>atory exercise seen as important to <strong>learning</strong> but not as a part <strong>of</strong> the „real <strong>work</strong>‟.<br />
<strong>The</strong> research in this thesis is guided by the idea that making retrospective <strong>reflection</strong><br />
integral to the <strong>work</strong> practice in project based <strong>learning</strong> can help students learn more from<br />
their project experience. <strong>The</strong> thesis explores how retrospective <strong>reflection</strong> in S<strong>of</strong>tware<br />
Engineering student projects can be supported, particularly taking into account the use<br />
<strong>of</strong> collaboration tools – typically lightweight tools – by teams in their daily project<br />
<strong>work</strong>.<br />
To this end, a set <strong>of</strong> interpretive case studies have been conducted. <strong>The</strong> studies examine<br />
the current use <strong>of</strong> state-<strong>of</strong>-the-art lightweight collaboration tools in S<strong>of</strong>tware<br />
Engineering student projects, focusing on the role <strong>of</strong> the tools in <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
within the teams <strong>and</strong> in collaboration with other project stakeholders. <strong>The</strong> thesis<br />
research also includes studies in which interventions have been made in the projects by<br />
introducing facilitated retrospective <strong>reflection</strong> activities with <strong>and</strong> without the aid <strong>of</strong><br />
collaboration tools. <strong>The</strong>se interventions can be considered as initial steps <strong>of</strong> a design<br />
research effort with the aim <strong>of</strong> improving the pedagogical design <strong>of</strong> the course as well<br />
as developing new theory on <strong>reflection</strong> in project based <strong>learning</strong>.<br />
<strong>The</strong> resulting contributions include new knowledge about the use <strong>of</strong> various types <strong>of</strong><br />
lightweight collaboration tools to support day-to-day <strong>work</strong> in the projects; within the<br />
teams <strong>and</strong> in collaboration between the teams <strong>and</strong> other project stakeholders. <strong>The</strong>se<br />
contributions can be used to aid the organization <strong>of</strong> future SE student projects including<br />
the choice <strong>and</strong> use <strong>of</strong> collaboration tools in the projects. <strong>The</strong> thesis also provides<br />
insights on how retrospective <strong>reflection</strong>, seen as a part <strong>of</strong> a collaborative <strong>work</strong> practice,<br />
can be supported in SE student projects <strong>and</strong> project <strong>work</strong> more generally. This research<br />
contribution includes a design for retrospective <strong>reflection</strong> <strong>work</strong>shops in educational<br />
settings (adapted from industry practice <strong>and</strong> empirically tested in a project course), a<br />
prototype tool extending wiki functionality to support navigation <strong>of</strong> relevant historical<br />
data, <strong>and</strong> empirical results demonstrating how a collaboration tool used in daily project<br />
<strong>work</strong> can be used as an aid to memory in retrospective <strong>reflection</strong>. Finally, a main<br />
contribution <strong>of</strong> this PhD <strong>work</strong> is a model <strong>of</strong> <strong>reflection</strong>, conceptualized in terms <strong>of</strong><br />
distributed cognition. Drawing on <strong>learning</strong> theory as well as current practice for<br />
S<strong>of</strong>tware Engineering retrospectives, the model represents a novel <strong>and</strong> practical view <strong>of</strong><br />
v
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
the <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in collaborative <strong>work</strong>. <strong>The</strong> model incorporates<br />
individual <strong>and</strong> collaborative steps <strong>of</strong> <strong>reflection</strong> on a collaborative <strong>work</strong> process, making<br />
explicit the role <strong>of</strong> internal <strong>and</strong> external representations <strong>of</strong> the process as an aid to<br />
<strong>reflection</strong>. <strong>The</strong> model also outlines the potential role <strong>of</strong> collaboration tools in supporting<br />
day-to-day <strong>work</strong> <strong>and</strong> retrospective <strong>reflection</strong> on that <strong>work</strong>, thereby integrating these<br />
aspects <strong>of</strong> the <strong>work</strong> practice. In this way shedding light on complexities <strong>and</strong><br />
opportunities related to <strong>reflection</strong> on collaborative <strong>work</strong> the model can aid design for<br />
retrospective <strong>reflection</strong> in project based <strong>learning</strong> in educational <strong>and</strong> pr<strong>of</strong>essional<br />
settings.<br />
vi
Preface<br />
This thesis is submitted to the Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
(NTNU) for partial fulfilment <strong>of</strong> the requirements for the degree <strong>of</strong> philosophiae doctor.<br />
<strong>The</strong> doctoral <strong>work</strong> has been performed at the <strong>Department</strong> <strong>of</strong> <strong>Computer</strong> <strong>and</strong> Information<br />
Science at NTNU. Pr<strong>of</strong>. Monica Divitini has been the main supervisor <strong>and</strong> Ove<br />
Haugaløkken has been the co-supervisor.<br />
<strong>The</strong> doctoral <strong>work</strong> is funded by MOTUS2, a research project within the NTNU research<br />
programme for Learning <strong>and</strong> ICT.<br />
vii
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
viii
Acknowledgements<br />
I would like to thank all the people who in different ways have contributed to this thesis.<br />
My colleagues at IDI have been an invaluable resource for discussing <strong>and</strong> exchanging<br />
ideas <strong>and</strong> have provided good company throughout my PhD years. Sobah Abbas<br />
Petersen <strong>and</strong> Kirsti Berntsen have been great <strong>of</strong>fice mates <strong>and</strong> friends. Glenn<br />
Munkvold, being a couple <strong>of</strong> years ahead <strong>of</strong> me on the PhD schedule, provided valuable<br />
support <strong>and</strong> advice on many aspects <strong>of</strong> PhD <strong>work</strong>. Torstein Hjelle is always helpful <strong>and</strong><br />
faithfully joins in on the morning visit to the c<strong>of</strong>fee shop. Thomas Østerlie, Sven<br />
Ziemer, Tor Stålhane, Torgeir Dingsøyr, Eric Monteiro, Dag Svanæs, Ilaria Canova<br />
Calori, Reidar Conradi, Gro Alice Hamre, Øyvind Hauge, Hallvard Trætteberg,<br />
Gasparas Jarulaitis, Letizia Jaccheri, Ole Andreas Alsos, Guttorm Sindre, Gry Sel<strong>and</strong>,<br />
Inger Dybdahl Sørby <strong>and</strong> Chiara Rossitto have all been generous in taking the time to<br />
discuss with me when I have been in need <strong>of</strong> opinions <strong>and</strong> insights. More names could<br />
have been added to this list, <strong>and</strong> I colloquially thank the people who make my daily<br />
<strong>work</strong> at IDI interesting <strong>and</strong> enjoyable.<br />
Thanks to Bendik Bygstad <strong>and</strong> Tor-Morten Grønli, my former colleagues at NITH, for<br />
good collaboration on research papers.<br />
Glenys Hamilton did efficient editing <strong>work</strong>, <strong>and</strong> I recommend her as an editor.<br />
Ove Haugaløkken, my co-supervisor, <strong>and</strong> Leif Martin Hokstad, in practice my second<br />
co-supervisor, have provided constructive <strong>and</strong> useful input from the side <strong>of</strong> educational<br />
theory <strong>and</strong> computer-supported collaborative <strong>learning</strong> throughout the PhD period.<br />
Special thanks go to my main supervisor Monica Divitini. She is supportive <strong>and</strong><br />
generous <strong>and</strong> has been doing a great job as co-author <strong>of</strong> several <strong>of</strong> my research papers.<br />
Applying her strong analytic skills to my ideas <strong>and</strong> writings she has frequently provided<br />
the essential keys to bringing the PhD <strong>work</strong> another step forward.<br />
Thanks to my parents Ellen <strong>and</strong> Jan Rognebakke <strong>and</strong> to my friends Idunn Løvseth<br />
Kavlie <strong>and</strong> Ragnhild Torsvik Bamrud for listening when I have been in need <strong>of</strong> talking<br />
about the challenges <strong>of</strong> life as a PhD student.<br />
Finally I thank my beloved children Øystein, Håvard <strong>and</strong> Idunn for putting up with a<br />
mother who has <strong>of</strong>ten been stressed, impatient <strong>and</strong> absorbed in her PhD <strong>work</strong>. I thank<br />
my dear husb<strong>and</strong> John for love <strong>and</strong> support. I am very fortunate to be married to a man<br />
who not only knows the requirements <strong>of</strong> academic <strong>work</strong> but who is also highly<br />
respectful <strong>of</strong> our differences in terms <strong>of</strong> coping with <strong>work</strong> <strong>and</strong> life.<br />
ix
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
x
Contents<br />
1 Introduction ........................................................................................................ 1<br />
1.1 Problem statement 1<br />
1.2 Problem refinement <strong>and</strong> research questions 3<br />
1.3 Research design 4<br />
1.4 Results <strong>and</strong> contributions 5<br />
1.5 Structure <strong>of</strong> the thesis 7<br />
2 Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical background ................................. 9<br />
2.1 Underst<strong>and</strong>ing <strong>work</strong> <strong>and</strong> <strong>learning</strong> 9<br />
2.1.1 A constructionist perspective on <strong>work</strong> <strong>and</strong> <strong>learning</strong> ............................ 9<br />
2.1.2 Communities <strong>of</strong> practice <strong>and</strong> <strong>learning</strong> ................................................ 10<br />
2.1.3 Distributed cognition .......................................................................... 12<br />
2.2 Reflection as a bridge between <strong>work</strong> <strong>and</strong> <strong>learning</strong> 13<br />
2.3 Trajectory as representation <strong>of</strong> a process 16<br />
3 S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-the-art .............................. 19<br />
3.1 <strong>The</strong> characteristics <strong>of</strong> SE student projects 19<br />
3.2 Use <strong>of</strong> lightweight collaboration tools in SE student projects 22<br />
3.3 Supporting <strong>reflection</strong> on SE <strong>work</strong> 23<br />
3.4 Using <strong>work</strong> tools as tools for <strong>learning</strong> 26<br />
4 Research approach <strong>and</strong> research process ...................................................... 29<br />
5 Results ................................................................................................................ 37<br />
5.1 P1: Cross-Community Collaboration <strong>and</strong> Learning in Customer-Driven<br />
S<strong>of</strong>tware Engineering Student Projects 37<br />
5.2 P2: Power through brokering. OSS participation in SE projects 38<br />
5.3 P3: Do‟s <strong>and</strong> Don‟ts <strong>of</strong> Instant Messaging in Students‟ Project Work 40<br />
5.4 P4: <strong>The</strong> wiki as an integrative tool in project <strong>work</strong> 40<br />
xi
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
5.5 P5: Using Project Wiki History to Reflect on the Project Process 42<br />
5.6 P6: Shared timeline <strong>and</strong> individual experience: Supporting retrospective<br />
<strong>reflection</strong> in student s<strong>of</strong>tware engineering teams 44<br />
5.7 P7: A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> utilizing<br />
historical data in collaborative tools 46<br />
5.8 P8: Supporting Reflection in S<strong>of</strong>tware Development with Everyday<br />
Working Tools 48<br />
6 Contributions <strong>and</strong> implications ....................................................................... 53<br />
6.1 Project <strong>work</strong> as cross-community collaboration 53<br />
6.2 Lightweight collaboration tools in project <strong>work</strong> 55<br />
6.3 Supporting retrospective <strong>reflection</strong> 57<br />
6.4 A model <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>reflection</strong> 60<br />
7 Evaluation ......................................................................................................... 63<br />
7.1 Evaluation <strong>of</strong> the contributions with respect to the research questions 63<br />
7.2 Evaluation <strong>of</strong> the contributions with respect to state-<strong>of</strong>-the-art <strong>of</strong> the SE<br />
Education, TEL <strong>and</strong> CSCW research fields 64<br />
7.3 Evaluation <strong>of</strong> the research process 67<br />
7.3.1 On the <strong>work</strong> in the thesis as interpretive research .............................. 67<br />
7.3.2 On the <strong>work</strong> in the thesis as design research ...................................... 72<br />
7.4 On the choice <strong>of</strong> theory 73<br />
8 Conclusion <strong>and</strong> further <strong>work</strong> .......................................................................... 77<br />
9 References.......................................................................................................... 81<br />
Glossary .................................................................................................................... 91<br />
Appendix A: Research papers ................................................................................ 95<br />
Research paper P1 97<br />
Research paper P2 107<br />
Research paper P3 119<br />
xii
Research paper P4 135<br />
Research paper P5 149<br />
Research paper P6 161<br />
Research paper P7 171<br />
Research paper P8 189<br />
Appendix B: Other research publications ........................................................... 211<br />
Practice-Based Learning as Mobile Learning: <strong>The</strong> Role <strong>of</strong> Boundary Objects 211<br />
Learning from achievement: scaffolding student projects in s<strong>of</strong>tware engineering<br />
212<br />
xiii
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
List <strong>of</strong> Figures<br />
Figure 1: Most used terms in the thesis (the research papers excluded); diagram<br />
generated in Wordle (http://www.wordle.net) ......................................................... iii<br />
Figure 2: <strong>The</strong> problem domain <strong>of</strong> the PhD thesis ............................................................. 3<br />
Figure 3: Research papers <strong>and</strong> the main topics <strong>of</strong> the research contributions .................. 7<br />
Figure 4: <strong>The</strong> model <strong>of</strong> experiential <strong>learning</strong> (Kolb et al. 1971) .................................... 14<br />
Figure 5: A model <strong>of</strong> the reflective process. From (Boud et al. 1985a) ......................... 15<br />
Figure 6: Timeline indicating when the data collection for the empirical studies took<br />
place ........................................................................................................................ 30<br />
Figure 7: Team A having a meeting at the customer's site. <strong>The</strong> customer (second from<br />
the right) has ordered pizza. ................................................................................... 33<br />
Figure 8: <strong>The</strong> whiteboard after a <strong>reflection</strong> <strong>work</strong>shop with a student team. (<strong>The</strong> photo<br />
has been manipulated to make the experience curves more visible in print). ........ 34<br />
Figure 9: Field notes. ...................................................................................................... 35<br />
Figure 10: Screen shot from a project wiki main page as the project approached final<br />
deadline. Note the strikethroughs <strong>of</strong> completed tasks in the to-do-list. ................ 41<br />
Figure 11: Wiki Walkthrough Tool: adding a tag to the current version <strong>of</strong> a page........ 43<br />
Figure 12: Snapshot <strong>of</strong> the whiteboard after the <strong>reflection</strong> <strong>work</strong>shop <strong>of</strong> one project<br />
team, showing the shared project timeline with the individual satisfaction curves.<br />
<strong>The</strong> numbers along Max’s curve indicate where he explained this experience to the<br />
rest <strong>of</strong> the team. From P6. ...................................................................................... 46<br />
Figure 13: Retrospective <strong>reflection</strong> as distributed cognition. From P7. ......................... 47<br />
Figure 14: Examination <strong>of</strong> historical data in Trac. From P8. ........................................ 49<br />
Figure 15: A model outlining how Trac supports project <strong>work</strong> <strong>and</strong> retrospective<br />
<strong>reflection</strong> on the project process in the case study <strong>of</strong> P8. From P8. ....................... 51<br />
Figure 16: Diagrams showing the researchers' interpretation <strong>of</strong> how Team F changed<br />
their story <strong>of</strong> their project in their <strong>reflection</strong> <strong>work</strong>shop.......................................... 71<br />
xiv
Figure 17: Support for <strong>learning</strong> in S<strong>of</strong>tware Engineering student projects as approached<br />
in this thesis ............................................................................................................ 77<br />
xv
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
List <strong>of</strong> Tables<br />
Table 1: Research questions for the PhD thesis ............................................................... 4<br />
Table 2: Empirical research activity vs. research questions <strong>and</strong> research papers .......... 31<br />
Table 3: Timetable <strong>of</strong> the retrospective <strong>reflection</strong> <strong>work</strong>shop conducted with each <strong>of</strong> the<br />
teams in a SE project course. From P6. .................................................................. 45<br />
Table 4: <strong>The</strong>ory informing the <strong>work</strong> presented in the research papers .......................... 68<br />
xvi
Abbreviations<br />
CSCL<br />
CSCW<br />
IS<br />
NITH<br />
NTNU<br />
OSS<br />
TEL<br />
<strong>Computer</strong>-Supported Collaborative Learning<br />
<strong>Computer</strong>-Supported Cooperative Work<br />
Information Systems<br />
Norwegian School <strong>of</strong> Information Technology<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
Open Source S<strong>of</strong>tware<br />
Technology Enhanced Learning<br />
xvii
1 Introduction<br />
For we live not in a settled <strong>and</strong> finished world, but in one which is going<br />
on, <strong>and</strong> where our main task is prospective, <strong>and</strong> where retrospect – <strong>and</strong><br />
all knowledge as distinct from thought is retrospect – is <strong>of</strong> value in the<br />
solidity, security, <strong>and</strong> fertility it affords our dealings with the future.<br />
1.1 Problem statement<br />
(Dewey 2008 (1909))<br />
This thesis is about project based <strong>learning</strong> (PBL) <strong>and</strong> how it can be supported in<br />
S<strong>of</strong>tware Engineering (SE) student projects. Project based <strong>learning</strong> is an educational<br />
approach at the intersection <strong>of</strong> <strong>work</strong> practice <strong>and</strong> education, <strong>and</strong> students‟ experience <strong>of</strong><br />
being practitioners is essential to their motivation <strong>and</strong> <strong>learning</strong> (Dohn 2009; Schön<br />
1987). <strong>The</strong> students do real <strong>work</strong> – independent, challenging, complex, aided by tools<br />
<strong>and</strong> in accordance with the norms <strong>and</strong> rules <strong>of</strong> the <strong>work</strong> practice, the latter mainly<br />
known to them through formal SE education. In project based <strong>learning</strong>, by participating<br />
in the community <strong>of</strong> practice (Wenger 1998) for which the project is preparing them,<br />
the students develop their insights <strong>and</strong> skills within their team as well as through<br />
interaction with other stakeholders such as the supervisor.<br />
Project activities perceived to be „school assignments‟ are not approached with the same<br />
enthusiasm as the challenges associated with doing „the real thing‟ in the project (i.e.<br />
solving the project task). In a SE student project, this task would typically be to develop<br />
a <strong>work</strong>ing s<strong>of</strong>tware product, maybe for an external customer.<br />
As pr<strong>of</strong>essional <strong>work</strong> practices involve collaboration (e.g. in project teams), project<br />
based <strong>learning</strong> typically takes place through <strong>work</strong> in teams. While collaboration adds<br />
opportunities to achieve what is not possible for individuals alone, it introduces<br />
complexity <strong>and</strong> the need for collaborative skills. In this thesis, focus will be on project<br />
based <strong>learning</strong> in teams.<br />
To support collaborative <strong>work</strong>, tools are used. In PBL, getting the desired <strong>work</strong><br />
experience includes getting experience with the relevant tools. Knowledge about<br />
adequate tool use can be drawn from the <strong>work</strong> domain in question (e.g. in the case <strong>of</strong> SE<br />
1
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
student projects, using tools similar to those used in SE industry). However, adding to<br />
the complexity <strong>of</strong> PBL, the adequate use <strong>of</strong> tools for supporting <strong>work</strong> may not be so<br />
clearly determined: as new technologies <strong>and</strong> patterns <strong>of</strong> using them rapidly develop<br />
across different domains, student teams may find themselves at the frontier <strong>of</strong><br />
technology use. For instance, current project <strong>work</strong> practices frequently involve the use<br />
<strong>of</strong> lightweight collaborative tools, many <strong>of</strong> which are being used for social as well as<br />
pr<strong>of</strong>essional purposes, <strong>and</strong> many <strong>of</strong> which are especially widespread among young<br />
people. <strong>The</strong> use <strong>of</strong> such tools in <strong>work</strong>place settings is currently being explored (Decker<br />
et al. 2007; Fuchs-Kittowski <strong>and</strong> Köhler 2005; Isaacs et al. 2002b; Stafford 2008). In<br />
the research community <strong>of</strong> <strong>Computer</strong> Supported Cooperative Work (CSCW), it is<br />
acknowledged that further research is needed in this area (Pipek <strong>and</strong> Wulf 2007).<br />
Learning from experience entails <strong>reflection</strong> (Boud et al. 1985b; Dewey 1933; Kolb <strong>and</strong><br />
Fry 1975). In context <strong>of</strong> a <strong>work</strong> practice, <strong>work</strong> <strong>and</strong> <strong>reflection</strong> are intertwined. Some<br />
<strong>reflection</strong> is inherent to the meaning making (Bruner 1990) <strong>of</strong> day-to-day <strong>work</strong>, <strong>and</strong><br />
some requires more distance to the experience reflected upon (Schön 1983). <strong>The</strong> aspects<br />
<strong>of</strong> <strong>work</strong> that can be elaborated through <strong>reflection</strong> range from those that are easily made<br />
explicit to those that can be considered tacit (Polanyi 1966) but may be codified <strong>and</strong><br />
shared within a common frame <strong>of</strong> reference (Hildrum 2009), <strong>and</strong> include the affective<br />
elements <strong>of</strong> personal experience (Boud et al. 1985b; Eraut 2000).<br />
<strong>The</strong> role <strong>of</strong> <strong>reflection</strong> in <strong>learning</strong> has long been recognized in the educational field <strong>and</strong><br />
incorporated into current practices (e.g. <strong>of</strong> project based <strong>learning</strong> (Duim et al. 2007;<br />
Helle et al. 2006)). In industry, the need to learn from experience is manifest in current<br />
industry practices <strong>of</strong> facilitated <strong>reflection</strong>, such as in project post-mortem evaluations<br />
(Dingsøyr 2005; Kerth 2001). Still, in educational as well as <strong>work</strong>place settings, there<br />
are obstacles to making <strong>reflection</strong> a useful part <strong>of</strong> the <strong>work</strong> practice that the <strong>reflection</strong> is<br />
intended to support (Armarego 2007; Kasi et al. 2008).<br />
Due to the central role <strong>of</strong> tools in mediating <strong>work</strong> <strong>and</strong> <strong>learning</strong> (Kirschner <strong>and</strong> Erkens<br />
2006; Kuutti <strong>and</strong> Kaptelinin 1997; Stahl 2002), research on the support for <strong>reflection</strong> in<br />
project based <strong>learning</strong> has a place in the research fields <strong>of</strong> Technology Enhanced<br />
Learning (TEL) <strong>and</strong> <strong>Computer</strong>-Supported Collaborative Learning (CSCL).<br />
As an empirical case <strong>of</strong> project based <strong>learning</strong> for the PhD <strong>work</strong> presented in this thesis,<br />
S<strong>of</strong>tware Engineering (SE) student projects were selected. <strong>The</strong> projects are<br />
collaborative, complex <strong>and</strong> based on active use <strong>of</strong> state-<strong>of</strong>-the-art collaboration<br />
technology. <strong>The</strong> extensive use <strong>of</strong> lightweight collaboration tools in the projects makes<br />
them a good case for exploring the role <strong>of</strong> this type <strong>of</strong> technology in <strong>work</strong> <strong>and</strong> <strong>learning</strong>.<br />
Regarding support for <strong>reflection</strong>, SE project course designs typically include activities<br />
2
Introduction<br />
to help the students reflect on their project process. However, there is a tendency that<br />
such activities are conducted only half-heartedly, being perceived as school assignments<br />
rather than part <strong>of</strong> the „real <strong>work</strong>‟. This indicates a <strong>learning</strong> potential left partially<br />
unexploited.<br />
How <strong>work</strong> <strong>and</strong> <strong>reflection</strong> are, <strong>and</strong> could be, supported <strong>and</strong> connected in SE student<br />
projects, taking into account the role <strong>of</strong> tools, is the problem domain (Figure 2) to be<br />
explored in this thesis. <strong>The</strong> domain is approached from different angles, framed in the<br />
research contexts <strong>of</strong> CSCW, TEL/CSCL <strong>and</strong> SE education.<br />
Figure 2: <strong>The</strong> problem domain <strong>of</strong> the PhD thesis<br />
1.2 Problem refinement <strong>and</strong> research questions<br />
<strong>The</strong> main research question for the PhD <strong>work</strong> is:<br />
How can <strong>work</strong> <strong>and</strong> <strong>reflection</strong> be supported <strong>and</strong> bridged in s<strong>of</strong>tware<br />
engineering student projects, taking into account the use <strong>of</strong><br />
collaboration tools?<br />
<strong>The</strong> refinement <strong>of</strong> the main research question into a set <strong>of</strong> research questions guiding<br />
the investigation <strong>of</strong> the problem domain was based on the following research strategy:<br />
• Start by examining current practices <strong>of</strong> <strong>work</strong> in student SE teams (particularly<br />
stressing the role <strong>of</strong> collaboration technology) <strong>and</strong> thereby identify challenges to<br />
<strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
• Provide solutions to the identified challenges, <strong>and</strong> evaluate these solutions.<br />
3
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
<strong>The</strong> following four research questions were formulated (Table 1):<br />
Table 1: Research questions for the PhD thesis<br />
RQ1: What characterizes s<strong>of</strong>tware engineering student projects?<br />
RQ2: What is the current usage <strong>of</strong> lightweight collaboration tools to support <strong>work</strong> in<br />
s<strong>of</strong>tware engineering student projects?<br />
RQ3: How can retrospective <strong>reflection</strong> be supported in s<strong>of</strong>tware engineering student<br />
projects?<br />
RQ4: How can collaboration tools that are used in daily project <strong>work</strong> be utilized as<br />
tools for project based <strong>learning</strong>?<br />
1.3 Research design<br />
Case studies <strong>of</strong> SE projects with a focus on the as-is situation were appropriate for<br />
unveiling current challenges in the projects in line with RQ1 <strong>and</strong> RQ2 <strong>and</strong> guide the<br />
selection <strong>of</strong> issues to be explored more deeply in the thesis. Particularly in the case <strong>of</strong><br />
two longitudinal studies, the research was guided by the principles <strong>of</strong> interpretive field<br />
studies (Klein <strong>and</strong> Myers 1999).<br />
To explore the selected issue <strong>of</strong> retrospective <strong>reflection</strong> in accordance with RQ3 <strong>and</strong><br />
RQ4, the research was designed to include interventions to current practice, exploring<br />
<strong>and</strong> evaluating approaches to retrospective <strong>reflection</strong>. This can be considered part <strong>of</strong> a<br />
design research (Kelly et al. 2008b) process with the goal <strong>of</strong> improving the SE project<br />
course in question as well as contributing with theoretical insights <strong>and</strong> practical<br />
approaches in the area <strong>of</strong> project based <strong>learning</strong> more generally.<br />
A SE project course at NTNU was a good c<strong>and</strong>idate for the empirical research needed<br />
to explore RQ1-4. <strong>The</strong> course is a capstone project course <strong>of</strong> an undergraduate IT study<br />
program, <strong>and</strong> while the students in the projects are relative novices in the field <strong>of</strong><br />
S<strong>of</strong>tware Engineering, they have relatively advanced SE knowledge <strong>and</strong> skills. Further<br />
adding to the realism <strong>of</strong> the projects as compared to pr<strong>of</strong>essional SE <strong>work</strong>, the projects<br />
have external customers for whom <strong>work</strong>ing s<strong>of</strong>tware products are to be developed.<br />
Being a researcher <strong>and</strong> some <strong>of</strong> the time course staff in various roles in the project<br />
course, I had opportunities to get insights about the course organization <strong>and</strong> about<br />
specific projects <strong>of</strong> interest. I had access to various project documentation <strong>and</strong><br />
opportunities for data collection across the teams in the course as well as the<br />
4
Introduction<br />
opportunity to do longitudinal studies <strong>of</strong> single teams. This permitted the collection <strong>of</strong> a<br />
large amount <strong>of</strong> data over several years to gain an in-depth underst<strong>and</strong>ing <strong>of</strong> the case.<br />
Also, I had the opportunity to make interventions with the course itself. <strong>The</strong> dual role <strong>of</strong><br />
researcher <strong>and</strong> staff <strong>and</strong> its possible impacts on the research were systematically<br />
addressed in the research. While most <strong>of</strong> the case material for the thesis was drawn from<br />
the SE project course at NTNU, some <strong>of</strong> the empirical data for research paper P1 was<br />
collected in a similar project course at NITH, the university college that was my former<br />
<strong>work</strong>place.<br />
In my PhD research I take a constructionist perspective on <strong>work</strong> <strong>and</strong> <strong>learning</strong>,<br />
deliberately seeking to focus on their social <strong>and</strong> cognitive aspects. I have chosen to be<br />
relatively restrictive in applying theoretical concepts in the initial phases <strong>of</strong> the research,<br />
including theory as seen useful for the different research activities to support data<br />
collection <strong>and</strong> analysis in a way coherent with a constructionist view.<br />
1.4 Results <strong>and</strong> contributions<br />
<strong>The</strong> research questions RQ1-RQ4 are answered in the following research papers:<br />
P1<br />
P2<br />
P3<br />
Krogstie, B. <strong>and</strong> B. Bygstad. Cross-Community Collaboration <strong>and</strong> Learning in<br />
Customer-Driven S<strong>of</strong>tware Engineering Student Projects. Twentieth Conference<br />
on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training (CSEE&T) 2007, Dublin.<br />
IEEE <strong>Computer</strong> Society. (Krogstie <strong>and</strong> Bygstad 2007)<br />
Krogstie, B. Power through brokering. OSS participation in SE projects. ICSE<br />
2008, Leipzig. IEEE <strong>Computer</strong> Society. (Krogstie 2008a)<br />
Krogstie, B.R. Do‟s <strong>and</strong> Don‟ts <strong>of</strong> Instant Messaging in Students‟ Project Work.<br />
NOKOBIT 2009, Trondheim. Tapir. (Krogstie 2009a)<br />
P4 Krogstie, B.R. <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. in COOP 2008.<br />
Carry-le-Rouet, Provence, France. Institut d‟Etudes Politiques d‟Aix-en-<br />
Provence. (Krogstie 2008b)<br />
P5<br />
P6<br />
Krogstie, B.R. Using Project Wiki History to Reflect on the Project Process.<br />
42nd Hawaii International Conference on System Sciences (HICSS‟42) 2009.<br />
Big Isl<strong>and</strong>, Hawaii: IEEE <strong>Computer</strong> Society. (Krogstie 2009c)<br />
Krogstie, B.R. <strong>and</strong> M. Divitini. Shared timeline <strong>and</strong> individual experience:<br />
Supporting retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering teams.<br />
CSEE&T 2009, Hyderabad. IEEE <strong>Computer</strong> Society. (Krogstie <strong>and</strong> Divitini<br />
2009)<br />
5
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
P7<br />
P8<br />
Krogstie, B.R. A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong><br />
utilizing historical data in collaborative tools. EC-TEL 2009, Nice. Springer.<br />
(Krogstie 2009b)<br />
Krogstie, B.R. <strong>and</strong> M. Divitini. Supporting Reflection in S<strong>of</strong>tware Development<br />
with Everyday Working Tools. COOP 2010, Aix-en-Provence, France. Springer.<br />
(Krogstie <strong>and</strong> Divitini 2010)<br />
In Figure 3 the four transparent rectangles indicate the general research topics addressed<br />
in the research contributions <strong>of</strong> the thesis <strong>and</strong> which research papers address each topic.<br />
<strong>The</strong> topics are:<br />
• Project <strong>work</strong> as cross-community collaboration (Section 6.1)<br />
• <strong>The</strong> use <strong>of</strong> lightweight collaboration tools (Section 6.2)<br />
• Retrospective <strong>reflection</strong>, with <strong>and</strong> without the aid <strong>of</strong> collaboration tools (Section<br />
6.3)<br />
• <strong>The</strong> integration <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>reflection</strong> in a <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong>,<br />
taking a distributed cognition perspective (Section 6.4)<br />
<strong>The</strong> „swim lanes‟ in the diagram show the research communities (SE Education, TEL<br />
<strong>and</strong> CSCW) in which the research papers have been peer reviewed <strong>and</strong> accepted for<br />
publication.<br />
6
Introduction<br />
Figure 3: Research papers <strong>and</strong> the main topics <strong>of</strong> the research contributions<br />
1.5 Structure <strong>of</strong> the thesis<br />
<strong>The</strong> thesis is structured as follows:<br />
Chapter 2 outlines relevant background theory on <strong>work</strong>, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>, with a<br />
focus on the choice <strong>of</strong> an adequate theoretical frame<strong>work</strong> to address the main research<br />
question <strong>of</strong> the thesis.<br />
Chapter 3 gives an overview <strong>of</strong> state-<strong>of</strong>-the art research in the areas <strong>of</strong> project based<br />
<strong>learning</strong> <strong>and</strong> lightweight tool support for <strong>work</strong> in S<strong>of</strong>tware Engineering <strong>and</strong> SE student<br />
projects.<br />
Chapter 4 presents the research approach <strong>of</strong> the PhD <strong>work</strong> <strong>and</strong> discusses its strengths<br />
<strong>and</strong> limitations.<br />
7
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Chapter 5 summarizes the results from the research papers.<br />
Chapter 6 outlines the contributions <strong>of</strong> the thesis, addressing their implications.<br />
Chapter 7 contains an evaluation <strong>of</strong> the PhD <strong>work</strong>.<br />
Chapter 8 concludes the thesis <strong>and</strong> suggests future research.<br />
Appendix A contains the PhD papers P1-P8.<br />
Appendix B contains the abstract <strong>of</strong> two research papers that were written during the<br />
PhD scholarship period but not included in the PhD paper collection<br />
8
2 Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical<br />
background<br />
[<strong>The</strong>ory] can be a useful signpost for those who need directions, or for<br />
that matter for those who are busily writing PhD theses <strong>and</strong> feel they need<br />
a crutch.<br />
(R<strong>and</strong>all et al. 2007)<br />
In this chapter, I briefly account for the theory guiding the research in the thesis.<br />
Concepts from various theoretical frame<strong>work</strong>s have been combined in order to shed<br />
light on issues <strong>of</strong> interest within a coherent conceptual frame. I start by summarizing<br />
theory on how to underst<strong>and</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong>. I next consider theory illuminating how<br />
<strong>reflection</strong> can bridge <strong>work</strong> <strong>and</strong> <strong>learning</strong>. Finally, I turn to the concept <strong>of</strong> trajectory as a<br />
way <strong>of</strong> conceiving <strong>and</strong> making sense <strong>of</strong> a process.<br />
2.1 Underst<strong>and</strong>ing <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
Outlining the main theory on <strong>work</strong> <strong>and</strong> <strong>learning</strong> underlying the thesis I address the<br />
constructionist perspective, theory <strong>of</strong> Communities <strong>of</strong> Practice (CoP) <strong>and</strong> <strong>learning</strong>, <strong>and</strong><br />
the theoretical frame<strong>work</strong> <strong>of</strong> Distributed cognition.<br />
2.1.1 A constructionist perspective on <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
Constructionism <strong>and</strong> constructivism are the core concepts <strong>of</strong> a large <strong>and</strong> non-unified<br />
body <strong>of</strong> research (Phillips 1995) building on the assumption that human knowledge is<br />
constructed in an interplay between individuals‟ thought processes <strong>and</strong> social processes<br />
(Berger <strong>and</strong> Luckmann 1966; Brown et al. 1989; Palincsar 1998; Vygotsky 1978). This<br />
view “discards the notion that knowledge could or should be a representation <strong>of</strong> an<br />
observer-independent world-in-itself <strong>and</strong> replaces it with the dem<strong>and</strong> that the conceptual<br />
constructs we call knowledge be viable in the experiential world <strong>of</strong> the knowing<br />
subject” (von Glasersfeld 1989). Such a perspective has pr<strong>of</strong>ound implications for how<br />
<strong>learning</strong> is regarded, <strong>and</strong> sets learners‟ experience at the centre <strong>of</strong> attention. <strong>The</strong><br />
9
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
constructionist perspective is common in CSCL research (Stahl 2002; Strijbos et al.<br />
2004; Suthers 2006) <strong>and</strong> also within CSCW (e.g. with activity theory (Kaptelinin <strong>and</strong><br />
Nardi 2006)).<br />
<strong>The</strong> constructionist perspective has implications for the thesis in the following ways:<br />
• Ontologically/epistemologically: <strong>The</strong> reality to be understood <strong>and</strong> designed for<br />
(in student SE project teams) should be considered a social construct.<br />
Participants (including the researcher) constantly engage in the construction <strong>of</strong><br />
this reality.<br />
• Research focus: If we want to underst<strong>and</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong>, knowledge<br />
construction among participants is an important object <strong>of</strong> study.<br />
• Research approach: It must always be kept in mind that participants have<br />
different interpretations <strong>and</strong> that these are made in certain social <strong>and</strong> historical<br />
contexts (Klein <strong>and</strong> Myers 1999).<br />
Viewing human activity mainly from a socio-cultural perspective or mainly from a<br />
cognitive one can be seen as two different str<strong>and</strong>s <strong>of</strong> constructionism. In this thesis, I<br />
take an intermediate position (Cobb 1994; Lin et al. 1999), stressing the social as well<br />
as the cognitive aspects <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong> in collaborative settings. This provides a<br />
starting point for shedding light on the interplay between thought processes <strong>and</strong> social<br />
processes.<br />
2.1.2 Communities <strong>of</strong> practice <strong>and</strong> <strong>learning</strong><br />
Communities <strong>of</strong> practice (CoP) (Lave <strong>and</strong> Wenger 1991; Wenger 1998) are an<br />
important arena for the construction <strong>of</strong> knowledge. CoPs are characterized by members<br />
having mutual engagement, a joint enterprise, <strong>and</strong> a shared repertoire. <strong>The</strong>y have a<br />
shared history <strong>of</strong> <strong>learning</strong> <strong>and</strong> boundaries with other communities. Learning the practice<br />
<strong>of</strong> a CoP is something that happens through participation in the community. A new<br />
participant entering a community starts out as a novice engaged in “legitimate<br />
peripheral participation” <strong>and</strong> gradually increases expertise through situated <strong>learning</strong><br />
(Lave <strong>and</strong> Wenger 1991). Along the same line, Brown <strong>and</strong> colleagues describe <strong>learning</strong><br />
as situated cognition (Brown et al. 1989), arguing that activity <strong>and</strong> situations are<br />
integral to cognition <strong>and</strong> <strong>learning</strong>.<br />
A student team engaged in a project throughout a semester can be seen as a CoP, albeit<br />
a small <strong>and</strong> relatively short-lived one. <strong>The</strong> team develops a shared history <strong>of</strong> <strong>learning</strong><br />
<strong>and</strong> has boundaries with other communities (e.g. those <strong>of</strong> project stakeholders). <strong>The</strong><br />
10
Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical background<br />
members <strong>of</strong> the teams are simultaneously members <strong>of</strong> other communities such as the<br />
pr<strong>of</strong>ession for which the education is preparing the students, <strong>and</strong> the students‟ personal<br />
circles <strong>of</strong> friends. <strong>The</strong>re are frequently different levels <strong>of</strong> expertise among the students<br />
with respect to the practice <strong>of</strong> the project. For instance, in the case <strong>of</strong> SE projects,<br />
programming skills <strong>of</strong>ten vary greatly within a team. From a CoP perspective, team<br />
members‟ <strong>work</strong> <strong>and</strong> <strong>learning</strong> thus takes place within <strong>and</strong> across communities with<br />
varying goals, norms <strong>and</strong> practices, <strong>and</strong> with expertise in these practices differing<br />
among community members <strong>and</strong> over time. In the thesis, team members‟ multiple<br />
memberships in different communities are a core issue in the research papers P1 <strong>and</strong> P2.<br />
A potential arena for <strong>learning</strong> in the context <strong>of</strong> CoPs is the boundaries between<br />
communities. Individuals who are simultaneously members <strong>of</strong> two communities can<br />
serve as brokers, contributing to the transfer <strong>and</strong> transformation <strong>of</strong> knowledge from one<br />
community into the other. This typically involves artifacts (Gal et al. 2005; Star 1990)<br />
integral to the practice <strong>of</strong> each <strong>of</strong> the communities <strong>and</strong> serving as boundary objects.<br />
Changes to the boundary objects (e.g. through collaboration), may have implications for<br />
the communities involved <strong>and</strong> be a source <strong>of</strong> <strong>learning</strong> <strong>and</strong> change. In the thesis, the<br />
issue <strong>of</strong> brokering is important in P2.<br />
In research papers P3 <strong>and</strong> P4, which focus on current use <strong>of</strong> lightweight technologies in<br />
SE student teams, the CoP concept helps shed light on the differences between activity<br />
within the team <strong>and</strong> that which involves collaboration with other stakeholders (e.g.<br />
customer, supervisor) <strong>and</strong> can be considered cross-community collaboration.<br />
A student SE team can also be considered as a particular type <strong>of</strong> CoP, namely a <strong>learning</strong><br />
community (Riel <strong>and</strong> Polin 2004). Riel <strong>and</strong> Polin provide a categorization <strong>of</strong> online<br />
<strong>learning</strong> communities which is also appropriate for characterizing communities <strong>of</strong><br />
learners that are fully or partially collocated, such as student SE team. <strong>The</strong> categories<br />
are intended by the authors to be analytic categories, <strong>and</strong> a specific <strong>learning</strong> community<br />
may have the characteristics <strong>of</strong> one or more <strong>of</strong> them:<br />
Task-based <strong>learning</strong> communities are “groups <strong>of</strong> people organized around a task<br />
who <strong>work</strong> intently together for a specified period <strong>of</strong> time to produce a product.<br />
While the specific group may not, in the strictest sense, share all <strong>of</strong> the<br />
properties <strong>of</strong> a community, the people who participate in them <strong>of</strong>ten experience<br />
a strong sense <strong>of</strong> identification with their partners, the task, <strong>and</strong> the organization<br />
that supports them.” (Riel <strong>and</strong> Polin 2004, p.20)<br />
Practice-based <strong>learning</strong> communities are larger groups with shared goals that<br />
<strong>of</strong>fer their members richly contextualized <strong>and</strong> supported arenas for <strong>learning</strong>.<br />
11
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Knowledge-based <strong>learning</strong> communities “<strong>of</strong>ten share many <strong>of</strong> the features <strong>of</strong> a<br />
community <strong>of</strong> practice but focus on the deliberate <strong>and</strong> formal production <strong>of</strong><br />
external knowledge about the practice”. This includes being committed to<br />
recording <strong>and</strong> sharing knowledge “outside its immediate use or active context”<br />
(Riel <strong>and</strong> Polin 2004, p.21).<br />
In SE student projects <strong>and</strong> project based <strong>learning</strong> more generally, while knowledge<br />
building beyond the project itself may be important in some projects, the teams are<br />
typical task-based <strong>learning</strong> communities. This theoretical point is central to the<br />
perspective on the use <strong>of</strong> lightweight tools in research paper P4. At the same time in<br />
project based <strong>learning</strong>, participation in authentic <strong>work</strong> (e.g the CoP <strong>of</strong> the pr<strong>of</strong>essional<br />
community or possibly a particular organization) is a motivating force. Issues arising<br />
from SE project students simultaneously being members <strong>of</strong> different types <strong>of</strong> <strong>learning</strong><br />
communities are discussed in P1.<br />
2.1.3 Distributed cognition<br />
From a constructionist perspective, <strong>learning</strong> in a SE student team can be viewed in<br />
terms <strong>of</strong> knowledge construction. To underst<strong>and</strong> how knowledge is constructed (e.g. by<br />
the aid <strong>of</strong> technology) it is useful to go into more detail on the role <strong>of</strong> tools <strong>and</strong> how<br />
knowledge is distributed.<br />
<strong>The</strong> impact on activity (e.g. project <strong>work</strong>) <strong>of</strong> tools is captured in the idea <strong>of</strong> mediation:<br />
the use <strong>of</strong> tools forming the activity <strong>and</strong> the activity over time forming the design <strong>of</strong> the<br />
tools. This is a key point in various theoretical frame<strong>work</strong>s explaining the interplay <strong>of</strong><br />
the individual <strong>and</strong> the social (e.g. activity theory (Engeström 1987; Leont'ev 1981;<br />
Vygotsky 1978) <strong>and</strong> actor net<strong>work</strong> theory (Latour 2005)). Tools can be understood in a<br />
broad sense, comprising physical artifacts, theories <strong>and</strong> concepts as well as<br />
computerized tools. Within the frame<strong>work</strong> <strong>of</strong> distributed cognition (DCog) (Hutchins<br />
1995), the central unit <strong>of</strong> analysis is the functional system, “a collection <strong>of</strong> individuals<br />
<strong>and</strong> artifacts <strong>and</strong> their relations to each other in a particular <strong>work</strong> practice” (Rogers <strong>and</strong><br />
Ellis 1994). As in actor net<strong>work</strong> theory, the artifacts (e.g. tools) are considered entities<br />
analytically on a par with human actors, which helps shed light on the dynamics <strong>of</strong> tool<br />
mediation. Analyzing a case <strong>of</strong> collaborative <strong>work</strong> from the perspective <strong>of</strong> DCog, the<br />
main goal is to account for how the distributed structures are coordinated. This implies<br />
examining the contributions <strong>of</strong> the environment in which the <strong>work</strong> activity takes place,<br />
individuals‟ interactions <strong>and</strong> the use <strong>of</strong> artifacts in interaction, <strong>and</strong> the representational<br />
media used. DCog has a focus on the way knowledge is propagated across different<br />
representational states along various communicative pathways (Rogers <strong>and</strong> Ellis 1994).<br />
12
Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical background<br />
In approaching a case <strong>of</strong> collaborative <strong>work</strong> with the aim <strong>of</strong> underst<strong>and</strong>ing the role <strong>of</strong><br />
tools in knowledge construction, the focus on representational media <strong>and</strong> states makes<br />
DCog a particularly useful analytical frame<strong>work</strong>. DCog has been used as a frame<strong>work</strong><br />
for underst<strong>and</strong>ing collaborative <strong>work</strong> in general (Rogers <strong>and</strong> Ellis 1994), organizational<br />
memory (Ackerman <strong>and</strong> Halverson 2004), SE project <strong>work</strong> (Sharp <strong>and</strong> Robinson 2006)<br />
as well as educational practice (Bl<strong>and</strong>ford <strong>and</strong> Furniss 2006; Daradoumis <strong>and</strong> Marques<br />
2002). In this thesis, DCog has been applied in the analysis <strong>of</strong> the case study in P8 <strong>and</strong><br />
in the development <strong>of</strong> the conceptual model <strong>of</strong> <strong>reflection</strong> presented in P7.<br />
A further link between tools <strong>and</strong> knowledge construction can be made by considering<br />
computerized tools that are used to aid thinking as mindtools (Pea 1985) or cognitive<br />
tools (Kim <strong>and</strong> Reeves 2007; Kirschner <strong>and</strong> Erkens 2006; Kuutti <strong>and</strong> Kaptelinin 1997).<br />
Cognitive tools can be seen as “cognitive <strong>reflection</strong> <strong>and</strong> amplification tools that help<br />
learners to construct their own realities by designing their own knowledge bases”<br />
(Jonassen 1995, p.43). In the thesis, the concept <strong>of</strong> cognitive tools is central to P7.<br />
<strong>The</strong> DCog perspective can shed light on the role <strong>of</strong> individual participants in the<br />
functional system <strong>of</strong> collaborative <strong>work</strong> by distinguishing between internal <strong>and</strong> external<br />
representations. <strong>The</strong> need to focus on the individual learner within a DCog perspective<br />
on collaborative <strong>learning</strong> has been stressed (Salomon 1993). From a related theoretical<br />
viewpoint, personal knowledge is “the cognitive resources which a person brings to a<br />
situation which enables them to think <strong>and</strong> perform” when (typically) combined with<br />
codified knowledge (Eraut 2000). <strong>The</strong> latter is by definition explicit, whereas personal<br />
knowledge can be either explicit or tacit (Polanyi 1966). In this thesis, the DCog<br />
perspective is used in P7 <strong>and</strong> P8 as the role <strong>of</strong> external <strong>and</strong> internal representations <strong>of</strong> a<br />
<strong>work</strong> process in <strong>work</strong> <strong>and</strong> <strong>reflection</strong> is explored.<br />
2.2 Reflection as a bridge between <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
Reflection is a core concept in the thesis <strong>and</strong> a main topic <strong>of</strong> the research papers P5-P8.<br />
Situated in the philosophical school <strong>of</strong> pragmatism, John Dewey developed a theory <strong>of</strong><br />
<strong>learning</strong> from experience:<br />
“To „learn from experience‟ is to make a backward <strong>and</strong> forward<br />
connection between what we do to things <strong>and</strong> what we enjoy or<br />
suffer from things in consequence. Under such conditions, doing<br />
becomes a trying; an experiment with the world to find out what it<br />
is like; the undergoing becomes instruction – discovery <strong>of</strong> the<br />
connection <strong>of</strong> things.” (Dewey 2008 (1909), p.140)<br />
13
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
To underst<strong>and</strong> the details <strong>of</strong> this connection, reflective thinking is needed. Dewey<br />
defined this as “active, persistent, <strong>and</strong> careful consideration <strong>of</strong> any belief or supposed<br />
form <strong>of</strong> knowledge in the light <strong>of</strong> the grounds that support it <strong>and</strong> the further conclusions<br />
in the direction <strong>of</strong> which it tends” (Dewey 1933, p.133).<br />
Kolb, Rubin <strong>and</strong> McIntyre (1971) presented a model <strong>of</strong> the <strong>learning</strong> process termed an<br />
experiential <strong>learning</strong> model, including a step <strong>of</strong> <strong>reflection</strong> (“Observations <strong>and</strong><br />
<strong>reflection</strong>s”) preceding the formation <strong>of</strong> abstract concepts <strong>and</strong> generalizations.<br />
Figure 4: <strong>The</strong> model <strong>of</strong> experiential <strong>learning</strong> (Kolb et al. 1971)<br />
Stahl (2002) presents a model <strong>of</strong> collaborative knowledge building which integrates<br />
<strong>cycle</strong>s <strong>of</strong> „building personal knowing‟ with a <strong>cycle</strong> <strong>of</strong> „building collaborative knowing‟.<br />
<strong>The</strong> model is fairly complex, including steps <strong>of</strong> externalizing collaborative knowing<br />
into cultural artifacts which again mediate activity <strong>and</strong> thereby individuals‟ personal<br />
knowing <strong>and</strong> contribution to the collective discourse <strong>and</strong> interaction. While Kolb places<br />
stronger focus on experience <strong>and</strong> Stahl on knowledge building, the models can both be<br />
seen as varieties <strong>of</strong> the <strong>learning</strong> <strong>cycle</strong>: the collective version <strong>of</strong> Kolb‟s „formation <strong>of</strong><br />
abstract concepts <strong>and</strong> generalizations‟ <strong>of</strong> which implications are to be tested in new<br />
situations (arguably) corresponds to Stahl‟s externalizing <strong>of</strong> collaborative knowing into<br />
cultural artifacts which again mediate activity.<br />
Experience is not purely rational, <strong>and</strong> not all experience can easily be expressed. This<br />
point is made by McCarthy <strong>and</strong> Wright (2004), addressing experience with technology:<br />
“Developing an account <strong>of</strong> felt experience with technology is difficult partially because<br />
the word „experience‟ is simultaneously rich <strong>and</strong> elusive. It is also difficult because we<br />
can never step out <strong>of</strong> experience <strong>and</strong> look at it in a detached way” (p.15).<br />
<strong>The</strong> complexity <strong>of</strong> experience <strong>and</strong> the necessity to consider more than its rational <strong>and</strong><br />
easily observable aspects are captured in the model <strong>of</strong> the reflective process (Figure 5)<br />
14
Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical background<br />
presented by Boud, Keogh <strong>and</strong> Walker (1985a; 1985b), who extend Kolb‟s <strong>learning</strong><br />
<strong>cycle</strong> <strong>and</strong> go further into details on the constituent elements <strong>of</strong> <strong>reflection</strong>.<br />
Figure 5: A model <strong>of</strong> the reflective process. From (Boud et al. 1985a)<br />
Experiences(s) is what is reflected upon, including behaviour, ideas <strong>and</strong> feelings. (One<br />
may for instance think <strong>of</strong> the experience <strong>of</strong> collaborating with others in a project over a<br />
period <strong>of</strong> time.) <strong>The</strong> reflective process consists <strong>of</strong> returning to experience, attending to<br />
feelings, <strong>and</strong> re-evaluating experience (e.g. considering what was good <strong>and</strong> bad about a<br />
project process). <strong>The</strong>is model indicates a possible iteration between experiencing <strong>and</strong><br />
reflecting. Outcomes include new perspectives on the experience reflected upon, change<br />
in behaviour, readiness for application, <strong>and</strong> commitment to action (e.g. to change an<br />
ongoing project in another direction or to make sure to do a certain task in a certain<br />
phase <strong>of</strong> the next project). <strong>The</strong> authors stress that an instance <strong>of</strong> <strong>reflection</strong> need not<br />
contain all the elements outlined in the model. Also, there is no fixed sequence <strong>of</strong> steps<br />
in the process. Still, with its checklist <strong>of</strong> elements <strong>of</strong> a reflective process <strong>and</strong> its<br />
grounding in <strong>learning</strong> theory, the model can be used both to underst<strong>and</strong> reflective<br />
processes <strong>and</strong> to aid the design for them. In the thesis, the model has been used in the<br />
research presented in P6 <strong>and</strong> P7.<br />
From a DCog perspective, <strong>reflection</strong> can be understood as an intrinsic aspect <strong>of</strong> how<br />
information is used (i.e. through simultaneous consideration <strong>of</strong> multiple representations,<br />
internal <strong>and</strong> external). Information is seen as interactive <strong>and</strong> “used cognitively <strong>and</strong><br />
socially. By interactive, it is meant that external representations used in situ are<br />
interpreted in coordination with the implicitly shared <strong>and</strong> individual knowledge <strong>of</strong><br />
current events <strong>and</strong> <strong>work</strong> practices” (Rogers <strong>and</strong> Ellis 1994, p.131). Bruner, in<br />
accounting for human meaning making (Bruner 1990), touches upon similar issues. He<br />
argues that meaning making is about negotiating <strong>and</strong> renegotiating by the mediation <strong>of</strong><br />
narrative interpretation: “ [] human beings, in interacting with another, form a sense <strong>of</strong><br />
the canonical <strong>and</strong> ordinary as a background against which to interpret <strong>and</strong> give narrative<br />
meaning to breaches in <strong>and</strong> deviations from „normal‟ states <strong>of</strong> the human condition”<br />
15
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
(Bruner 1990, p.67). <strong>The</strong> comparison <strong>of</strong> different representations can be seen as a key<br />
point in <strong>reflection</strong>. For instance, a project team‟s <strong>reflection</strong> on their project process may<br />
entail the team members‟ consideration <strong>and</strong> interpretation <strong>of</strong> an external representation<br />
<strong>of</strong> the process created by one team member, or the team‟s collaboratively constructed<br />
timeline <strong>of</strong> project events considered in light <strong>of</strong> the „ideal‟ process from the point <strong>of</strong><br />
view <strong>of</strong> a certain methodology for conducting the project <strong>work</strong>.<br />
Many aspects <strong>of</strong> <strong>work</strong> experience worth <strong>learning</strong> from, <strong>and</strong> thus worth returning to in a<br />
reflective process, are tacit (Polanyi 1966) <strong>and</strong> not immediately accessible to thought<br />
<strong>and</strong> discussion. Hildrum argues that sharing <strong>of</strong> some tacit knowledge through<br />
codification is possible between participants who have a shared frame <strong>of</strong> reference<br />
(Hildrum 2009). From the perspective <strong>of</strong> social <strong>learning</strong> theory (B<strong>and</strong>ura 1996) <strong>learning</strong><br />
from tacit knowledge can happen through observational <strong>learning</strong>, which includes<br />
observation <strong>and</strong> elaboration <strong>of</strong> symbolic models <strong>of</strong> the experience. <strong>The</strong>se are examples<br />
<strong>of</strong> research on <strong>learning</strong> that can be used to underpin the idea that <strong>reflection</strong>, by aid <strong>of</strong><br />
appropriate external representations (e.g. as seen within DCog), can help participants<br />
learn from aspects <strong>of</strong> <strong>work</strong> that are in themselves less explicit. In the thesis, project<br />
team members‟ externalization <strong>of</strong> knowledge about their project process as an important<br />
part <strong>of</strong> <strong>reflection</strong>, is a main concern in P6, P7 <strong>and</strong> P8.<br />
Finally, <strong>reflection</strong> as part <strong>of</strong> a <strong>work</strong> practice is to a large extent integral to daily <strong>work</strong>,<br />
taking the form <strong>of</strong> <strong>reflection</strong>-in-action. Sometimes <strong>reflection</strong> requires more distance<br />
from the activity, which can be seen as <strong>reflection</strong>-on-action. (Schön 1983; Schön 1987).<br />
this thesis, while not diminishing the significance <strong>of</strong> <strong>reflection</strong>-in-action, has a focus on<br />
<strong>reflection</strong> with some distance to the process reflected upon, looking particularly at<br />
facilitated retrospective <strong>reflection</strong> (P5-P8).<br />
2.3 Trajectory as representation <strong>of</strong> a process<br />
Strauss (1993) argues that when people make sense <strong>of</strong> a process (e.g. one <strong>of</strong> cooperative<br />
<strong>work</strong>), they think <strong>of</strong> it as a trajectory. A trajectory is “(1) the course <strong>of</strong> any experienced<br />
phenomenon as it evolves over time” <strong>and</strong> “(2) the actions <strong>and</strong> interactions contributing<br />
to its evolution” (pp.53-54). Trajectories can be individual or shared, <strong>and</strong> they can be<br />
past ones or projected future ones. <strong>The</strong> role <strong>of</strong> trajectories as seen by Strauss bears<br />
resemblance to the role <strong>of</strong> narratives in meaning making (Bruner 1990).<br />
Strauss‟ trajectory concept is part <strong>of</strong> his theory <strong>of</strong> action, a frame<strong>work</strong> for underst<strong>and</strong>ing<br />
human acting <strong>and</strong> social phenomena, influenced by pragmatism (e.g. (Dewey 1978<br />
(1910))) <strong>and</strong> interactionism (e.g. (Blumer 1986; Mead 1934)). <strong>The</strong> theory <strong>of</strong> action was<br />
used by Fitzpatrick in the Locales frame<strong>work</strong> (Fitzpatrick 2003), illustrating the<br />
16
Work, <strong>learning</strong> <strong>and</strong> <strong>reflection</strong>: theoretical background<br />
usefulness <strong>of</strong>, for example, the trajectory concept in the context <strong>of</strong> underst<strong>and</strong>ing<br />
cooperative <strong>work</strong> <strong>and</strong> support for that <strong>work</strong>.<br />
<strong>The</strong> basic assumptions about social interaction underlying Strauss‟ theory <strong>of</strong> action are<br />
compatible with the theoretical underpinnings <strong>of</strong> the <strong>work</strong> in this thesis. <strong>The</strong> trajectory<br />
concept has been applied in the research papers P5-P8 as a core concept shedding light<br />
on how team members conceptualize a project process <strong>and</strong> what may be adequate<br />
external representations <strong>of</strong> a project process when people (e.g. project team members)<br />
are to reconstruct, discuss <strong>and</strong> reflect on their process.<br />
17
3 S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-theart<br />
In this section, I provide a brief overview <strong>of</strong> state-<strong>of</strong>-the-art research identifying<br />
challenges that are addressed through the research contributions <strong>of</strong> the thesis. <strong>The</strong><br />
chapter is organized in line with the research questions RQ1-RQ4<br />
3.1 <strong>The</strong> characteristics <strong>of</strong> SE student projects<br />
A first main characteristic <strong>of</strong> SE student projects is that they are instances <strong>of</strong> s<strong>of</strong>tware<br />
engineering. <strong>The</strong> IEEE <strong>Computer</strong> Society defines s<strong>of</strong>tware engineering as: "(1) <strong>The</strong><br />
application <strong>of</strong> a systematic, disciplined, quantifiable approach to the development,<br />
operation, <strong>and</strong> maintenance <strong>of</strong> s<strong>of</strong>tware; that is, the application <strong>of</strong> engineering to<br />
s<strong>of</strong>tware. (2) <strong>The</strong> study <strong>of</strong> approaches as in (1)." (IEEE 1990). <strong>The</strong> knowledge areas<br />
associated with s<strong>of</strong>tware engineering (Abran et al. 2004) comprise technical knowledge<br />
as well as knowledge <strong>of</strong> the human <strong>and</strong> social aspects <strong>of</strong> cooperation. S<strong>of</strong>tware<br />
development constitutes “an exercise in complex interrelationships” with frequent adhoc<br />
collaboration (Cherry <strong>and</strong> Robillard 2008). <strong>The</strong>re may be many project stakeholders<br />
(e.g. customers, users, technology providers), each with their own perspectives <strong>and</strong><br />
interests (Poole 2003). Collaboration with the stakeholders is essential <strong>and</strong> frequently<br />
challenging (Brady et al. 2006). <strong>The</strong> complexity <strong>of</strong> design <strong>work</strong> (Carstensen <strong>and</strong><br />
Schmidt 2002 (1999)) makes SE <strong>work</strong> unpredictable <strong>and</strong> difficult to coordinate. <strong>The</strong><br />
many project failures in SE industry (Keil et al. 2000) point to a need to learn from<br />
experience (Lyytinen <strong>and</strong> Robey 1999). <strong>The</strong> SE student projects <strong>of</strong> the case studies in<br />
this thesis have external customers, which contributes to a realistic complexity <strong>of</strong> <strong>work</strong>.<br />
Based on recent increased underst<strong>and</strong>ing <strong>of</strong> the need for flexibility <strong>and</strong> adaptability to<br />
change in many SE projects, many popular s<strong>of</strong>tware development methodologies are<br />
based on iterations <strong>and</strong> incremental development (Boehm 1988; Kruchten 2003) <strong>and</strong> –<br />
in case <strong>of</strong> agile methodologies – an increased focus on user/customer interaction rather<br />
than extensive documentation (Beck 2000; Cockburn 2006; Dybå <strong>and</strong> Dingsøyr 2008)<br />
(Germain <strong>and</strong> Robillard 2004). In a given s<strong>of</strong>tware development project, the choice <strong>of</strong><br />
development methodology should be made with the specific characteristics <strong>of</strong> that<br />
project in mind (Cockburn 2000). Project management is to some extent linked to the<br />
19
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
development methodology but is not necessarily strongly supported within the<br />
development methodology itself. In this thesis, the SE student projects <strong>of</strong> the case<br />
studies generally use iterative methodologies to structure their <strong>work</strong>.<br />
At this point there is a need for clarification about the use <strong>of</strong> the terms s<strong>of</strong>tware<br />
development (SD) <strong>and</strong> s<strong>of</strong>tware engineering (SE), for which definitions vary in the<br />
associated pr<strong>of</strong>essions <strong>and</strong> research fields. This thesis concentrates on the SD<br />
components <strong>of</strong> s<strong>of</strong>tware engineering <strong>work</strong>. In the research papers I use the term<br />
s<strong>of</strong>tware engineering when I refer to <strong>and</strong> address the research fields <strong>and</strong> pr<strong>of</strong>essions <strong>of</strong><br />
s<strong>of</strong>tware engineering <strong>and</strong> SE education. In the publications for the CSCW <strong>and</strong> TEL<br />
communities, the focus is on s<strong>of</strong>tware development as a particular domain <strong>of</strong> <strong>work</strong>, <strong>and</strong><br />
I refer to s<strong>of</strong>tware development. In the thesis introduction, to be consistent I use only the<br />
term s<strong>of</strong>tware engineering.<br />
A second, major characteristic <strong>of</strong> SE student projects is that they are based on the<br />
pedagogical approach <strong>of</strong> project based <strong>learning</strong> (PBL). To be considered an instance <strong>of</strong><br />
PBL, a project must have the following characteristics (Thomas 2000):<br />
• be central, not peripheral to the curriculum<br />
• be focused on questions or problems that “drive” students to encounter (<strong>and</strong><br />
struggle with) the central concepts <strong>and</strong> principles <strong>of</strong> a discipline<br />
• involve students in a constructive investigation<br />
• be student-driven to some significant degree<br />
• be realistic, not like school <strong>work</strong><br />
PBL bears similarity to problem based <strong>learning</strong>, <strong>and</strong> the distinction between the two<br />
approaches is not necessarily clean in practice (Prince <strong>and</strong> Felder 2007). One main<br />
difference is that PBL, as opposed to problem based <strong>learning</strong>, involves the construction<br />
<strong>of</strong> concrete artifacts (Helle et al. 2006). Another difference is that PBL is based on<br />
students mainly applying knowledge that is previously acquired, with the final product<br />
as the primary focus. Problem based <strong>learning</strong> generally has a stronger focus on the<br />
solution process <strong>and</strong> no formal instruction to provide the necessary background for<br />
solving the problem (Prince <strong>and</strong> Felder 2007).<br />
As argued in Section 1.1, PBL resides at the intersection between education <strong>and</strong> <strong>work</strong>.<br />
This creates a tension between the associated educational <strong>and</strong> practice logics (Dohn<br />
2009). Educational practices are generally based on an acquisition metaphor <strong>of</strong><br />
<strong>learning</strong>, in which the goals are external to participation: the learner, by participating in<br />
20
S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-the-art<br />
<strong>learning</strong> practice, should change in a way enabling more competent participation in<br />
<strong>work</strong>ing life practices. In other words, the educational practice should mainly prepare<br />
the learner for future participation in <strong>work</strong> life. At the same time, PBL is meant to<br />
provide <strong>work</strong> practice, the goals <strong>of</strong> which are intrinsic to participation (e.g. being a<br />
practitioner, proving oneself as a competent practitioner), solving the task being the<br />
main pro<strong>of</strong> <strong>of</strong> successful participation, <strong>and</strong> knowledge <strong>and</strong> competence being viewed as<br />
situated doing. Dohn (2009) argues that PBL is one <strong>of</strong> the pedagogical approaches that<br />
succeed in bridging the educational <strong>and</strong> practice logics by challenging the problem from<br />
within, “bringing authentic <strong>work</strong>ing life problems into schools settings as authentic<br />
problems (not just as illustrative examples) <strong>and</strong>/or by organizing educational tasks<br />
within <strong>work</strong>ing life settings.” (p.354).<br />
To provide support for PBL, it is useful to know how the students perceive the tension<br />
between education <strong>and</strong> <strong>work</strong>, <strong>and</strong>, in the same vein, what are their objectives for project<br />
participation. For instance, Berglund <strong>and</strong> Eckerdal (2006) found in a case study that<br />
project students strive from three different motives: academic achievement, project <strong>and</strong><br />
team <strong>work</strong>ing capacity (including becoming a better pr<strong>of</strong>essional), <strong>and</strong> social<br />
competence. Students‟ priorities among various objectives, <strong>and</strong> how they see certain<br />
tasks <strong>and</strong> aspects <strong>of</strong> project <strong>work</strong> as related to these objectives, have an impact on their<br />
<strong>work</strong> <strong>and</strong> <strong>learning</strong>, <strong>and</strong> ultimately on what are the most useful ways <strong>of</strong> supporting <strong>work</strong><br />
<strong>and</strong> <strong>learning</strong> in the projects.<br />
Also, students‟ underst<strong>and</strong>ing <strong>of</strong> the practices <strong>and</strong> objectives <strong>of</strong> the various CoPs<br />
involved (e.g. the university, the project team, the customer organization, the <strong>work</strong><br />
pr<strong>of</strong>ession) impacts on their collaboration with project stakeholders. Due to the<br />
importance <strong>and</strong> challenges <strong>of</strong> stakeholder collaboration in S<strong>of</strong>tware Engineering, it is<br />
acknowledged as vital to have SE project students learn to communicate with project<br />
stakeholders (McMillan 1999; Poole 2003).<br />
In this thesis, SE project students‟ objectives for project participation <strong>and</strong> underst<strong>and</strong>ing<br />
<strong>of</strong> the objectives <strong>of</strong> other stakeholders (e.g. the university, the customer organization)<br />
are explored in research paper P1. Collaboration between project teams <strong>and</strong><br />
stakeholders is addressed in the papers P2 <strong>and</strong> P3: P2 elaborating on a successful case<br />
<strong>of</strong> project-critical <strong>and</strong> tool-mediated collaboration, <strong>and</strong> P3 discussing an unsuccessful<br />
one.<br />
21
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
3.2 Use <strong>of</strong> lightweight collaboration tools in SE student<br />
projects<br />
In project based <strong>learning</strong>, tools are used to support <strong>learning</strong> <strong>and</strong> <strong>work</strong>. Technology<br />
introduced to support <strong>learning</strong>, what may be called <strong>learning</strong> technology, primarily has a<br />
role within the education logics <strong>of</strong> PBL. Learning technology includes tools for<br />
structuring <strong>learning</strong> activities (Xu 2007), having course staff communicate with,<br />
monitor <strong>and</strong> evaluate students (Coates et al. 2005; Trentin 2009) or in other ways<br />
support students‟ collaborative <strong>learning</strong> activities (e.g. building knowledge (Cress <strong>and</strong><br />
Kimmerle 2008), engaging in discussions (Zuhrieh 2009) <strong>and</strong> comparison with expert<br />
thinking (Lin et al. 1999)). <strong>The</strong> use <strong>of</strong> s<strong>of</strong>tware tools to support learners‟ construction <strong>of</strong><br />
knowledge representations to aid collaborative <strong>learning</strong> was addressed in Suthers <strong>and</strong><br />
Hundhausen (2001). Among the lightweight tools wikis are cited as particularly good<br />
for collaborative knowledge construction <strong>and</strong> open, low-threshold participation<br />
(Hickerson <strong>and</strong> Giglio 2009; Kim <strong>and</strong> Lee 2002; Lund <strong>and</strong> Smørdal 2006).<br />
While acknowledging the relevance <strong>of</strong> <strong>learning</strong> technology for PBL, this thesis has a<br />
focus on tools primarily used to support <strong>work</strong>, e.g. participation in SE practice. It is the<br />
role <strong>and</strong> usage <strong>of</strong> a tool in a project that makes it a <strong>work</strong> tool: for instance a wiki may be<br />
introduced by a SE team as a tool for project management.<br />
An important part <strong>of</strong> effective project communication is to use appropriate technology<br />
to support it (Burnett 2001). This points to the need for students in PBL not only to gain<br />
experience with using collaboration technologies, for example in stakeholder<br />
communication, but to the need to learn to judge the appropriateness <strong>of</strong> the technology<br />
for the purpose.<br />
<strong>The</strong> focus in the thesis is on the use <strong>of</strong> lightweight collaboration tools. <strong>The</strong>se are tools<br />
that can be acquired <strong>and</strong> used at low cost (e.g. money <strong>and</strong> time to learn) for individuals<br />
<strong>and</strong> their organization. A lightweight collaboration tool typically provides a limited set<br />
<strong>of</strong> features to support one aspect <strong>of</strong> collaborative <strong>work</strong> <strong>and</strong> may thus be relatively easily<br />
integrated into existing <strong>work</strong> processes rather than imposing a certain process on the<br />
user. Many lightweight collaboration tools are associated with Web 2.0 (Dohn 2009),<br />
for example wikis, discussion forums, <strong>and</strong> instant messaging.<br />
Lightweight collaboration tools are actively used in <strong>work</strong> life. Tools used to support<br />
<strong>work</strong> practices include blogs, instant messaging, discussion forums <strong>and</strong> wikis (e.g.<br />
Isaacs et al. 2002a; Lovejoy <strong>and</strong> Grudin 2003; Majchrzak et al. 2006; Muller et al.<br />
2003; Nardi et al. 2000; Niinimaki 2008; Quan-Haase et al. 2005). In the area <strong>of</strong><br />
S<strong>of</strong>tware Engineering, textual chat <strong>and</strong> discussion forums have been found to be<br />
22
S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-the-art<br />
powerful tools in distributed s<strong>of</strong>tware development (Gutwin et al. 2004a; Gutwin et al.<br />
2004b; Herbsleb et al. 2001). Wikis have been found useful for supporting various<br />
aspects <strong>of</strong> SE <strong>work</strong> (Louridas 2006; Radziwill <strong>and</strong> Shelton 2004) including stakeholder<br />
participation in requirements engineering (Decker et al. 2007).<br />
Lightweight collaboration tools that are used in pr<strong>of</strong>essional <strong>work</strong> are also used in other<br />
areas <strong>of</strong> life <strong>and</strong>/or among young people in particular (Baron 2004; Grinter <strong>and</strong> Palen<br />
2002; Huang <strong>and</strong> Yen 2003).<br />
In the thesis P3 provides a brief overview <strong>of</strong> the research literature on the use <strong>of</strong> instant<br />
messaging in project <strong>work</strong>, <strong>and</strong> P4 <strong>and</strong> P5 outlines relevant research on the use <strong>of</strong> wikis<br />
in SE project <strong>work</strong> as well as in student projects in particular.<br />
As lightweight collaboration technology continuously develops along with their patterns<br />
<strong>of</strong> use, there is a need for research investigating how these tools support the activities<br />
for which they are used (e.g. in s<strong>of</strong>tware engineering <strong>and</strong> project based <strong>learning</strong>). This<br />
thesis looks into current usage in SE project teams <strong>of</strong> the following tools: internet<br />
discussion forums (P2), instant messaging (P3), project wikis (P4) <strong>and</strong> issue trackers (a<br />
type <strong>of</strong> lightweight project management tool) (P8). Also, the concerted use <strong>of</strong> tools,<br />
including email, is addressed in several <strong>of</strong> the papers <strong>and</strong> is particularly central to the<br />
main argumentation <strong>of</strong> P7.<br />
Social web spaces like Facebook <strong>and</strong> MySpace are being actively used among students,<br />
<strong>and</strong> researchers have recently begun looking into the potential <strong>of</strong> such tools to support<br />
<strong>learning</strong> (Durkee et al. 2009; Idris <strong>and</strong> Wang 2009). From the empirical studies<br />
underlying this thesis there is no indication, informally assessed, that these tools are<br />
currently taken into use in the SE student projects to support project-related<br />
collaboration, or that other use <strong>of</strong> social web spaces has any significant impact on the<br />
project <strong>work</strong>. <strong>The</strong> use <strong>of</strong> social web spaces was defined to be outside the scope <strong>of</strong> this<br />
thesis.<br />
3.3 Supporting <strong>reflection</strong> on SE <strong>work</strong><br />
In educational contexts, the need to support <strong>reflection</strong> on <strong>learning</strong> processes (e.g. in<br />
project <strong>work</strong>) has been addressed by traditional approaches such as the use <strong>of</strong> <strong>reflection</strong><br />
notes <strong>and</strong> diaries (Gleaves et al. 2007). <strong>Computer</strong>ized tools designed to support<br />
<strong>reflection</strong> have been developed, some focusing on individual action <strong>and</strong> some on social<br />
discourse, with different ways <strong>of</strong> recording <strong>and</strong> representing the <strong>reflection</strong> <strong>and</strong> its<br />
subject matter (Kim <strong>and</strong> Lee 2002). A more recent example is the „challenges<br />
assessment <strong>work</strong>space‟ developed for project groups doing distributed team<strong>work</strong> (Xiao<br />
et al. 2008). After each project phase, the team had a structured assignment to<br />
23
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
collaboratively identify <strong>and</strong> evaluate (using shared spreadsheets <strong>and</strong> whiteboards) the<br />
challenges <strong>of</strong> team<strong>work</strong> in that phase <strong>of</strong> their project. <strong>The</strong> <strong>work</strong> <strong>of</strong> Xiao <strong>and</strong> colleagues<br />
exemplifies the use <strong>of</strong> a structured approach to <strong>reflection</strong> with the aid <strong>of</strong> a defined<br />
procedure, a shared <strong>work</strong>space <strong>and</strong> specified external representations.<br />
Turning to project <strong>work</strong> in industry, „lessons learned‟ – key project experiences with<br />
some general business relevance for future projects – has recently been incorporated as<br />
a concept in the Project Management Book <strong>of</strong> Knowledge (PMBOK)<br />
(ProjectManagementInstitute 2008). This points to a general acknowledgement in the<br />
project management field that <strong>learning</strong> from project experience needs to focus on a<br />
useful outcome. Further, to identify lessons learned from a project process, it is not only<br />
the „what‟, „where‟ <strong>and</strong> „how many‟ questions that need to be addressed, but also the<br />
„why‟ <strong>and</strong> „how‟ questions (Schindler <strong>and</strong> Eppler 2003). To achieve this it may be<br />
necessary to tap into participants‟ experience to try to learn from the tacit aspects <strong>of</strong><br />
project <strong>work</strong> (Eraut <strong>and</strong> Hirsch 2007).<br />
Project debriefings are structured approaches to <strong>reflection</strong> intended to help projects <strong>and</strong><br />
organizations capture <strong>and</strong> use experience data from projects (Schindler <strong>and</strong> Eppler<br />
2003). Schindler <strong>and</strong> Eppler examined a large body <strong>of</strong> empirical data on project-centred<br />
<strong>learning</strong> in various types <strong>of</strong> organizations. <strong>The</strong>y found that current practices, as<br />
compared to current theoretical insights, require “significant improvements with regard<br />
to the format, process, <strong>and</strong> use <strong>of</strong> lessons learned”. Schindler <strong>and</strong> Eppler list the<br />
following success factors for gaining from lessons learned in project debriefing<br />
<strong>work</strong>shops: 1) Have the entire project team regularly capture important project<br />
experiences after major milestones, 2) Have an external, neutral moderator <strong>of</strong> the<br />
debriefing <strong>work</strong>shops, 3) Aid the collection <strong>and</strong> structuring <strong>of</strong> lessons learned by a<br />
graphical representation in the form <strong>of</strong> a timeline <strong>and</strong> document the <strong>work</strong>shop in a<br />
poster format visible for all involved 4) Ensure that individual experiences are evaluated<br />
<strong>and</strong> analyzed in collective <strong>and</strong> interactive sessions, <strong>and</strong> 5) Gain commitment from<br />
gathered insights in the form <strong>of</strong> action consequences (i.e. what to do <strong>and</strong> by whom).<br />
In the S<strong>of</strong>tware Engineering industry, the prevalence <strong>and</strong> cost <strong>of</strong> project overruns <strong>and</strong><br />
failures (Keil et al. 2000) means the motivation to have teams <strong>and</strong> organizations learn<br />
from experience is strong. Efforts to achieve such <strong>learning</strong> may be integrated into<br />
predefined <strong>work</strong> processes <strong>and</strong> supported by tool functionality for collecting <strong>and</strong><br />
providing experience data. <strong>The</strong> idea <strong>of</strong> experience factories (Basili <strong>and</strong> Caldiera 1995)<br />
reflects a similar idea. In other settings, as in agile s<strong>of</strong>tware development (Dybå <strong>and</strong><br />
Dingsøyr 2008), <strong>learning</strong> from experience may be more informal, taking place in faceto-face<br />
project meetings <strong>and</strong> retrospective <strong>reflection</strong> sessions. Project retrospectives<br />
(Derby et al. 2006; Kerth 2001) in agile SE, conducted at the end <strong>of</strong> project phases or<br />
24
S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-the-art<br />
when projects are terminated, are an example <strong>of</strong> deliberate efforts to learn from<br />
experience, formally integrated into the <strong>work</strong> process (i.e. the development<br />
methodology). <strong>The</strong> major resources are participants‟ memory <strong>and</strong> collaborative effort to<br />
reconstruct the project. Various facilitation techniques exist (Bjørnson et al. 2009;<br />
Dingsøyr 2005) to help project participants gain an underst<strong>and</strong>ing <strong>of</strong> the „facts <strong>and</strong><br />
feelings‟ <strong>of</strong> their project process (thus approaching its tacit aspects), identify <strong>and</strong><br />
prioritize challenges <strong>and</strong> decide on how to address them.<br />
Providing support for <strong>work</strong> <strong>and</strong> <strong>learning</strong> based on representations <strong>of</strong> the <strong>work</strong> processes<br />
is a core idea in fields related to the management <strong>of</strong> <strong>work</strong>, for example in organizational<br />
<strong>learning</strong> <strong>and</strong> knowledge management (Walsham 2001). In project based <strong>learning</strong>,<br />
support for <strong>learning</strong> can to some extent be based on scaffolding (Wood et al. 1976) with<br />
the aid <strong>of</strong> pre-defined approaches (such as those outlined in formally defined s<strong>of</strong>tware<br />
development process models (Bygstad et al. 2009; Rooij 2009)). Lin <strong>and</strong> colleagues<br />
(Lin et al. 1999) point to four different ways <strong>of</strong> providing support for learners‟<br />
awareness about their <strong>learning</strong> process: process displays, showing learners explicitly<br />
what they are doing by making „normally tacit <strong>learning</strong> processes explicit <strong>and</strong> overt‟<br />
(p.47); process prompting (making students explain <strong>and</strong> evaluate their actions before,<br />
during or after problem-solving); process modelling focusing on how an expert would<br />
approach the problem; <strong>and</strong> finally reflective social discourse where the individual gets<br />
feedback from the community <strong>and</strong> accordingly modifies the practices. <strong>The</strong> last category<br />
includes using collaboratively constructed, external representations <strong>of</strong> the <strong>learning</strong><br />
subject matter, to support students‟ <strong>and</strong> teachers‟ sense making <strong>of</strong> the <strong>learning</strong> process<br />
(Pea 1994).<br />
In project based <strong>learning</strong> with a high level <strong>of</strong> independence in organizing the projects,<br />
<strong>and</strong> with project tasks that are not clearly defined in advance, the possibility to prespecify<br />
the desired <strong>work</strong> or <strong>learning</strong> process is limited. Reflection <strong>and</strong> <strong>learning</strong> from<br />
experience must to a large extent be based on participants‟ collaborative construction <strong>of</strong><br />
knowledge about their process <strong>and</strong> results. This can be achieved by the use <strong>of</strong> state-<strong>of</strong>the-art<br />
project retrospective techniques. In the thesis, research papers P6-P8 describe<br />
studies in which a project retrospective technique was adapted to, <strong>and</strong> applied in, a SE<br />
project course. <strong>The</strong> timeline <strong>and</strong> satisfaction curve approach (Derby et al. 2006; Kerth<br />
2001) supports the capturing <strong>and</strong> visualizing <strong>of</strong> participants‟ project experience,<br />
including its affective aspects. <strong>The</strong> focus <strong>of</strong> P5 is also on retrospective <strong>reflection</strong> on the<br />
project process, but in this case the <strong>reflection</strong> is facilitated by post-mortem interviews.<br />
25
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
3.4 Using <strong>work</strong> tools as tools for <strong>learning</strong><br />
In project <strong>work</strong> in industry, lessons learned are frequently not elicited <strong>and</strong> documented.<br />
Among the reasons are a lack <strong>of</strong> integration <strong>of</strong> experience recording into the project<br />
processes <strong>and</strong> team members believing that coding their experience is ineffective <strong>and</strong>/or<br />
not personally useful to them (Schindler <strong>and</strong> Eppler 2003). In the S<strong>of</strong>tware Engineering<br />
field, Kasi <strong>and</strong> colleagues (Kasi et al. 2008) conducted a Delphi study to find what<br />
experienced IT practitioners see as the most important barriers to conducting postmortem<br />
evaluations. Among the types <strong>of</strong> challenges to successful post-mortem<br />
evaluations found in the study was the lack <strong>of</strong> adequate data. „Hard data‟ about a project<br />
can help focus discussion in post-mortem evaluations (Collier et al. 1996). <strong>The</strong>se<br />
findings point to the significance <strong>of</strong> capturing project data that is relevant for team<br />
members‟ <strong>learning</strong>, <strong>and</strong> the need to facilitate <strong>reflection</strong> by having participants construct<br />
knowledge representations in a way perceived as feasible <strong>and</strong> meaningful.<br />
Information relevant to a <strong>work</strong> process can be captured from tools <strong>and</strong> artifacts already<br />
used to support that <strong>work</strong> process. <strong>The</strong> use <strong>of</strong> logs from a code management system to<br />
support coordination in a SD team has been proposed (Fitzpatrick et al. 2006). In that<br />
study, an event notification system <strong>and</strong> a tickertape tool for CVS (Concurrent Versions<br />
System, a popular version control system) messages was integrated with the code<br />
management system, improving communication in the team. <strong>The</strong> project memory tool<br />
Hipikat was designed to help new members <strong>of</strong> a distributed SD team learn from more<br />
experienced colleagues (Cubranic et al. 2004). <strong>The</strong> memory contains the artifacts<br />
produced during development <strong>and</strong> is built by the tool with little or no changes to the<br />
existing <strong>work</strong> practices. Tools for systematically capturing design rationale have been<br />
developed, e.g. IBIS (Conklin <strong>and</strong> Yakemovic 1991) <strong>and</strong> SEURAT (Burge <strong>and</strong> Brown<br />
2008). SEURAT captures design rationale in s<strong>of</strong>tware development <strong>and</strong> can also trace<br />
requirements <strong>and</strong> decisions (Burge <strong>and</strong> Kiper 2008) in a project. (Omoronyia et al.<br />
2009) outline an approach in which awareness support for s<strong>of</strong>tware development is<br />
provided through an awareness model derived from data originating in developers‟<br />
interaction in the shared <strong>work</strong>space. Some approaches to capturing data from a <strong>work</strong><br />
process involve users‟ deliberate tagging (performed as an integral part <strong>of</strong> the <strong>work</strong><br />
process) <strong>of</strong> information relevant to others (Dekel <strong>and</strong> Herbsleb 2008; Storey et al.<br />
2005).<br />
SE student projects represent a <strong>work</strong> <strong>and</strong> <strong>learning</strong> setting in which tool usage is<br />
dominated by lightweight technology. As part <strong>of</strong> their use in day-to-day <strong>work</strong>, these<br />
tools capture data (e.g. in logs, archives <strong>and</strong> various types <strong>of</strong> overviews) that may be <strong>of</strong><br />
use to aid <strong>reflection</strong> by the teams on their project process. <strong>The</strong> potential <strong>of</strong> collaboration<br />
tools to aid project <strong>reflection</strong> through team members‟ examination <strong>of</strong> historical data can<br />
26
S<strong>of</strong>tware Engineering student projects: state-<strong>of</strong>-the-art<br />
be examined in an informal way by having team members browse historical data as part<br />
<strong>of</strong> a post-mortem interview about their project. Such a study is reported in P5.<br />
A recent study found that in SD projects, data repositories from the development<br />
process (bug databases <strong>and</strong> email repositories) could help researchers reconstruct the<br />
bug fixing „stories‟, but the data were not complete <strong>and</strong> correct enough to get the full<br />
story deemed necessary to underst<strong>and</strong> <strong>and</strong> support coordination <strong>of</strong> the SD <strong>work</strong> (Ar<strong>and</strong>a<br />
<strong>and</strong> Venolia 2009). One <strong>of</strong> the biggest problems with the repository data was missing<br />
links from the bug records to the source code involved. Also, even with a bug database<br />
<strong>and</strong> email repositories in combination, it turned out to be difficult to unveil the full<br />
net<strong>work</strong> <strong>of</strong> participants involved in the process, <strong>and</strong> clearly see who had been doing<br />
what. <strong>The</strong> researchers got the full story <strong>of</strong> the bug fixing, with its social, organizational<br />
<strong>and</strong> technical aspects, by eliciting the knowledge <strong>of</strong> the participants in interviews.<br />
<strong>The</strong> use <strong>of</strong> historical data in collaboration tools to support <strong>reflection</strong> can be explored by<br />
incorporating such use <strong>of</strong> data into <strong>reflection</strong> efforts organized as part <strong>of</strong> the <strong>work</strong><br />
practice. This is in line with the pedagogical rationale <strong>of</strong> project based <strong>learning</strong>. In this<br />
thesis, the empirical study presented in P7 <strong>and</strong> P8 combines the timeline <strong>and</strong><br />
satisfaction curve project retrospective approach with the use <strong>of</strong> historical data in a<br />
collaboration tool (i.e. an issue tracker) to investigate what may be the benefit <strong>of</strong><br />
introducing the tool in this setting. P5 also explores the use <strong>of</strong> historical data in<br />
retrospective <strong>reflection</strong>, but framed within a tool-mediated post-mortem interview about<br />
the project.<br />
In the research literature on <strong>learning</strong> there are many models outlining <strong>learning</strong> <strong>cycle</strong>s<br />
from an educational or organizational perspective. Some <strong>of</strong> these models assign a key<br />
role to knowledge representations (see Section 2.2). However, to the knowledge <strong>of</strong> this<br />
author, none <strong>of</strong> these models outline the use <strong>of</strong> collaboration tools applied in the <strong>work</strong><br />
process as collectors <strong>and</strong> sources <strong>of</strong> data about the <strong>work</strong> process to be used in <strong>reflection</strong><br />
on that process. With the aid <strong>of</strong> a distributed cognition perspective, this challenge is<br />
undertaken in P7 <strong>and</strong> P8 in the thesis.<br />
27
4 Research approach <strong>and</strong> research process<br />
I write impressions during the research, after each interview, for example.<br />
I generate more organized sets <strong>of</strong> themes <strong>and</strong> issues after a group <strong>of</strong><br />
interviews or a major field visit. I then try to think about what I have learnt<br />
so far from my field data. If this sounds a rather subjective <strong>and</strong> relatively<br />
unplanned process, well it is. I believe that the researcher‟s best tool for<br />
analysis is his or her own mind, supplemented by the minds <strong>of</strong> others when<br />
<strong>work</strong> <strong>and</strong> ideas are exposed to them.<br />
(Walsham 2006)<br />
Simply observing <strong>learning</strong> <strong>and</strong> cognition as they naturally occur in the<br />
world is not adequate given that <strong>learning</strong> scientists frequently have<br />
transformative agendas<br />
(Barab <strong>and</strong> Squire 2004)<br />
Design experiments are conducted to develop theories, not merely to<br />
empirically tune „what <strong>work</strong>s‟<br />
(Cobb et al. 2003)<br />
<strong>The</strong> research <strong>of</strong> this thesis combines the use <strong>of</strong> interpretive case studies (Klein <strong>and</strong><br />
Myers 1999; Walsham 2006) <strong>and</strong> design research (Kelly 2003). As research approaches,<br />
they have in common a focus on trying to underst<strong>and</strong> specific settings; in the case <strong>of</strong><br />
design research with explicit intent to make an improvement. Interpretive case studies<br />
can be used to build a thorough underst<strong>and</strong>ing <strong>of</strong> the setting to be designed for.<br />
<strong>The</strong> combination <strong>of</strong> interpretive <strong>and</strong> design research in the thesis reflects an intent to do<br />
exploratory studies in an early phase <strong>of</strong> the <strong>work</strong> <strong>and</strong>, on the basis <strong>of</strong> relevant theory<br />
<strong>and</strong> empirically identified challenges, try out possible solutions <strong>and</strong> use the results to<br />
contribute to theory <strong>and</strong> practice.<br />
29
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
<strong>The</strong> two longitudinal field studies <strong>of</strong> single project teams (P2, P7) draw on naturalistic<br />
observation <strong>and</strong> the collection <strong>of</strong> a wide array <strong>of</strong> qualitative data <strong>and</strong> can be considered<br />
as ethnomethodologically informed (R<strong>and</strong>all et al. 2007).<br />
For the studies presented in P1 <strong>and</strong> P3-P5 data has been collected across the teams <strong>of</strong><br />
project courses (all teams, or a selection based, for example, on the use <strong>of</strong> a specific<br />
collaboration technology). <strong>The</strong> data collection was based on semi-structured interviews,<br />
various types <strong>of</strong> project documentation <strong>and</strong> access to logs from collaboration tools.<br />
<strong>The</strong> last part <strong>of</strong> the research involved tool design <strong>and</strong> intervention into the organization<br />
<strong>of</strong> a project course, with the objective <strong>of</strong> answering research questions RQ3 <strong>and</strong> RQ4. A<br />
prototype tool was developed, extending current functionality in a wiki tool (P4). It was<br />
tested mainly through expert evaluation <strong>of</strong> scenarios <strong>of</strong> use. Interventions to the project<br />
course were made with the aim <strong>of</strong> trying out approaches to retrospective <strong>reflection</strong> (P6-<br />
P8) at the end <strong>of</strong> the projects. Seen as a step towards improving the organization <strong>of</strong> the<br />
project course <strong>and</strong> informing the implementation <strong>of</strong> similar approaches to retrospective<br />
<strong>reflection</strong> in other project based <strong>learning</strong> settings, this research can be considered as<br />
initial steps <strong>of</strong> a design research process.<br />
This chapter gives a chronological account <strong>of</strong> the research process. An evaluation <strong>of</strong> the<br />
research as interpretive research <strong>and</strong> design research is provided in Chapter 7.<br />
I start by briefly explaining my pr<strong>of</strong>essional role with respect to the institutions in<br />
which the research took place. In the period 2000-2004 I was affiliated with NITH as<br />
academic staff, being responsible for, among other things, teaching <strong>and</strong> supervision in<br />
project courses. At NTNU, starting my <strong>work</strong> as a PhD student in 2005, I was course<br />
staff for the project course IT2901 in 2006 (as project supervisor), 2007 (as course<br />
coordinator) <strong>and</strong> 2008 (as course coordinator in the first half <strong>of</strong> the semester).<br />
Figure 6: Timeline indicating when the data collection for the empirical studies<br />
took place<br />
30
Research approach <strong>and</strong> research process<br />
<strong>The</strong> diagram in Figure 6 gives an overview <strong>of</strong> the major research activities <strong>of</strong> the PhD<br />
<strong>work</strong> (to be elaborated shortly) by showing how the data collection for the various<br />
studies was distributed in time.<br />
A more detailed overview <strong>of</strong> the research activities, showing their connection to the<br />
research questions <strong>and</strong> publications, is given in Table 2. <strong>The</strong> table also outlines which<br />
activities were carried out in the different educational institutions (NITH <strong>and</strong> NTNU).<br />
Table 2: Empirical research activity vs. research questions <strong>and</strong> research papers<br />
Institution<br />
/ project<br />
course<br />
NITH/<br />
PJ201<br />
NTNU/<br />
IT2901<br />
Year <strong>of</strong><br />
empirical<br />
study<br />
Participants Research activity RQs Research<br />
paper<br />
2006 7 <strong>of</strong> 11 teams Interviews 1-2 P1, P7<br />
6 <strong>of</strong> 14<br />
customers<br />
Interviews + analysis <strong>of</strong><br />
customer evaluation<br />
1<br />
All customers 1<br />
Team V + After-the-project 1-2 P3, P7<br />
customer interview + analysis <strong>of</strong><br />
project documentation<br />
2007 Team A + Longitudinal study 1-2 P2, P7<br />
project<br />
stakeholders<br />
4 teams (those<br />
with project<br />
wiki)<br />
All teams<br />
After-the-project<br />
interviews + analysis <strong>of</strong><br />
project documentation<br />
Interviews + analysis <strong>of</strong><br />
project documentation<br />
2008 Development <strong>of</strong> wiki<br />
walkthrough tool<br />
All teams Interviews + analysis <strong>of</strong><br />
project documentation<br />
Team F + Longitudinal study +<br />
project retrospective <strong>reflection</strong><br />
stakeholders <strong>work</strong>shop<br />
All teams Retrospective <strong>reflection</strong><br />
<strong>work</strong>shops<br />
1-2-<br />
3-4<br />
P4, P5,<br />
P7<br />
1-2 P3, P5<br />
3-4 P5, P7<br />
1-2 P3<br />
1-2-<br />
3-4<br />
P8, P7<br />
3 P6, P7<br />
As can be seen from Figure 6 <strong>and</strong> Table 2, the collection <strong>of</strong> empirical data took place in<br />
the period 2006-2008. <strong>The</strong> research papers were published in the period 2007-2010.<br />
<strong>The</strong> PhD research activity in 2005 mainly involved getting into the CSCW field <strong>and</strong><br />
included teaching a CSCW graduate course.<br />
31
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
In 2006, empirical studies were conducted at NITH <strong>and</strong> at NTNU. A longitudinal study<br />
<strong>of</strong> a NITH team involving some days <strong>of</strong> recorded observation over a semester was<br />
useful because it informed the focus <strong>of</strong> further research, but the study itself did not<br />
produce results that were published. <strong>The</strong> study reported in P1 addressed crosscommunity<br />
collaboration in capstone SE projects with external customers, drawing on<br />
data from semi-structured interviews with project students <strong>and</strong> customers in the<br />
undergraduate SE project course PJ501 at NITH, <strong>and</strong> interviews with customers from<br />
the NTNU project course IT2901 which was found to be similar enough to allow a<br />
combination <strong>of</strong> data across the courses. <strong>The</strong> study was conducted in collaboration with<br />
Bendik Bygstad at NITH.<br />
Team V at NTNU was identified as a particularly interesting case <strong>of</strong> instant messaging<br />
(IM) use, experiencing highly problematic team-customer collaboration over this<br />
medium. Based on the IM log <strong>and</strong> follow-up interviews with the project manager <strong>and</strong><br />
the customer, the findings were reported in P3.<br />
Data from the 2006 studies also informed a theoretical <strong>reflection</strong> paper (Krogstie <strong>and</strong><br />
Divitini 2007) on how the projects can be seen as a case <strong>of</strong> mobile <strong>learning</strong>, the<br />
scaffolding <strong>of</strong> which should focus on boundary objects between interacting<br />
communities (see Appendix B).<br />
From 2007 on, all data were collected from the project course IT2901 at NTNU. A<br />
longitudinal field study was conducted with Team A (Figure 7). It was made clear to the<br />
team that I would have no role in the evaluation <strong>of</strong> their <strong>work</strong> <strong>and</strong> provide no<br />
supervision. <strong>The</strong> study involved 75 hours <strong>of</strong> naturalistic observation <strong>of</strong> <strong>work</strong> sessions<br />
<strong>and</strong> meetings with project stakeholders. Data collection included field notes, photos <strong>and</strong><br />
audio recordings from the observed <strong>work</strong> sessions, the team‟s email communication,<br />
artifacts involved in the project <strong>work</strong>, project documentation at various stages <strong>of</strong><br />
development, follow-up interviews with the customer <strong>and</strong> the team right after the<br />
project, <strong>and</strong> two team members reading through a draft version <strong>of</strong> P2. Brokering<br />
towards an OSS development community was identified as a topic for P2, <strong>and</strong> as part <strong>of</strong><br />
the analysis the field notes were used to create a chronology <strong>of</strong> project events.<br />
32
Research approach <strong>and</strong> research process<br />
Figure 7: Team A having a meeting at the customer's site. <strong>The</strong> customer (second<br />
from the right) has ordered pizza.<br />
Data on the use <strong>of</strong> collaboration tools were collected across all teams in the 2007 cohort<br />
<strong>of</strong> the project course through mid-semester interviews <strong>and</strong> through inspection <strong>of</strong> project<br />
deliverables, particularly the <strong>reflection</strong> notes (for which a template had been made to<br />
make sure that the use <strong>of</strong> collaboration technology was addressed). <strong>The</strong>se data were<br />
later used in P3 <strong>and</strong> P7.<br />
Four 2007 teams were using project wikis, <strong>and</strong> three used them actively throughout<br />
their project. Data on wiki usage was collected by examining the wikis <strong>and</strong> interviewing<br />
students from the three active teams after project completion. Results on the use <strong>of</strong><br />
wikis in daily <strong>work</strong> were published in P4. Two <strong>of</strong> the retrospective interviews were<br />
aided by the examination <strong>of</strong> wiki contents, a form <strong>of</strong> mediation turning out to be so<br />
useful for project participants‟ recall <strong>and</strong> <strong>reflection</strong> that it was made a research topic<br />
informing the analysis <strong>of</strong> the interviews. <strong>The</strong> results were presented in P5.<br />
Pursuing the potential <strong>of</strong> wikis to aid retrospective <strong>reflection</strong> <strong>and</strong> the possibility to<br />
improve existing wiki functionality in this respect, I took the role <strong>of</strong> customer for a<br />
project team (Team W) in the spring semester <strong>of</strong> 2008. Team W developed the Wiki<br />
Walkthrough Tool (WWT), a prototype tool implemented as an add-on to an existing<br />
wiki tool (see Section 5.5) <strong>and</strong> presented in P4. <strong>The</strong> tool was tested in Team W‟s own<br />
project wiki in the last phase <strong>of</strong> their project. Additionally, use scenarios for the tool<br />
were evaluated by two expert teams, one <strong>of</strong> former project customers <strong>and</strong> one <strong>of</strong> project<br />
supervisors.<br />
In 2008, data on the use <strong>of</strong> collaboration tools in the teams were again collected through<br />
mid-semester interviews with all teams as well as examination <strong>of</strong> the teams‟ <strong>reflection</strong><br />
notes. <strong>The</strong> results were used in P3 <strong>and</strong> P7. Continuing the <strong>work</strong> on ways to support<br />
students‟ <strong>reflection</strong> on the project process, I introduced retrospective <strong>reflection</strong><br />
33
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
<strong>work</strong>shops at the end <strong>of</strong> the semester, adapting the timeline <strong>and</strong> satisfaction curve<br />
approach from SE industry (Figure 8) (see 5.6). I was the <strong>work</strong>shop facilitator,<br />
„external‟ by having no involvement with the evaluation <strong>of</strong> the projects <strong>and</strong> keeping the<br />
outcomes confidential apart from anonymous use in research. <strong>The</strong> results were<br />
published in P6.<br />
Figure 8: <strong>The</strong> whiteboard after a <strong>reflection</strong> <strong>work</strong>shop with a student team. (<strong>The</strong><br />
photo has been manipulated to make the experience curves more visible in print).<br />
Throughout the spring semester 2008, a longitudinal study was made <strong>of</strong> Team F. <strong>The</strong><br />
approach was similar to the one used for the study <strong>of</strong> Team A in 2007. I had good<br />
access to project artifacts <strong>and</strong> the developing status <strong>of</strong> the project <strong>work</strong>, being updated<br />
through the team‟s email list <strong>and</strong> by looking into the issue tracker (Trac) used to<br />
manage the development <strong>work</strong>. Sample field notes are shown in Figure 9. During<br />
observation sessions, field notes were made on a PC. Pictures were taken <strong>and</strong> included<br />
directly in the electronic document as needed. After the sessions, the notes were cut <strong>and</strong><br />
pasted into a paper notebook <strong>and</strong> supplemented with h<strong>and</strong>written comments.<br />
A retrospective <strong>reflection</strong> <strong>work</strong>shop was conducted with Team F at the end <strong>of</strong> the<br />
semester. This was an adapted version <strong>of</strong> the timeline <strong>and</strong> satisfaction curve approach<br />
used in the <strong>work</strong>shops for all the other teams in the course. In the case <strong>of</strong> team F, the<br />
use <strong>of</strong> historical data in Trac was incorporated into the <strong>work</strong>shop design to see if<br />
participants could use the tool to recall project events they had not recalled by memory<br />
alone, <strong>and</strong> if this had an impact on their <strong>reflection</strong>. <strong>The</strong> <strong>work</strong>shop lasted 3 days<br />
altogether including individual <strong>and</strong> collective sessions (for details, see P7). <strong>The</strong><br />
<strong>work</strong>shop was video recorded in a usability lab, providing synchronized screen<br />
recordings <strong>of</strong> the use <strong>of</strong> Trac <strong>and</strong> the team members‟ activity in the room. Analysis <strong>of</strong><br />
the <strong>work</strong>shop data acknowledged background knowledge <strong>of</strong> the team‟s project process<br />
gained through the longitudinal study. Results were published in P7 <strong>and</strong> P8. P7 details<br />
34
Research approach <strong>and</strong> research process<br />
the use <strong>of</strong> Trac in the context <strong>of</strong> <strong>reflection</strong> (e.g. important tool features). In P8, the<br />
results are generalized to a model <strong>of</strong> <strong>reflection</strong> incorporating the use <strong>of</strong> historical data in<br />
collaboration tools.<br />
Figure 9: Field notes.<br />
35
5 Results<br />
This chapter presents an overview <strong>of</strong> the results paper by paper. For each paper, the<br />
original abstract is included, followed by a short description <strong>of</strong> the background for the<br />
research <strong>and</strong> a brief elaboration on the results.<br />
5.1 P1: Cross-Community Collaboration <strong>and</strong> Learning in<br />
Customer-Driven S<strong>of</strong>tware Engineering Student Projects<br />
Authors: Krogstie, Birgit <strong>and</strong> Bygstad, Bendik<br />
I was the first author <strong>of</strong> this paper. Both authors participated in the research design, data<br />
collection, analysis <strong>and</strong> writing process.<br />
Published in: Proceedings <strong>of</strong> the 20th Conference on S<strong>of</strong>tware Engineering Education<br />
<strong>and</strong> Training (CSEE&T) 2007<br />
This paper explores collaboration <strong>and</strong> <strong>learning</strong> between stakeholders in customer-driven student<br />
projects. <strong>The</strong> research objectives are to obtain empirically based knowledge on how students relate to<br />
stakeholders in customer-driven projects, <strong>and</strong> to suggest implications for the pedagogical design <strong>of</strong> the<br />
project courses. Empirical data was collected from two Bachelor courses in s<strong>of</strong>tware engineering at two<br />
<strong>learning</strong> institutions in Norway. To make sense <strong>of</strong> the interaction between the three stakeholders in the<br />
project: the student groups, the university <strong>and</strong> the customer, we build on Wenger’s concept <strong>of</strong><br />
communities <strong>of</strong> practice <strong>and</strong> on the concept <strong>of</strong> boundary objects. Our analysis highlights that students,<br />
through practical experience in the projects, learn to balance the requirements <strong>and</strong> expectations from<br />
different stakeholders in designing a <strong>work</strong>ing technical solution - a valuable skill for s<strong>of</strong>tware engineers.<br />
We argue that for students to learn to balance stakeholders’ interests in the best possible way, visibility <strong>of</strong><br />
stakeholders’ goals should be focused throughout the projects. Explicit reference to the goals should be<br />
incorporated into project artifacts serving as boundary objects. Collaboration technologies providing<br />
st<strong>and</strong>ard shared <strong>work</strong>space functionality are seen as adequate to support this.<br />
Based on the authors‟ <strong>work</strong> experience with capstone SE projects as well as the data<br />
collected in all teams <strong>of</strong> a capstone project course in 2006, the importance <strong>and</strong><br />
mechanisms <strong>of</strong> cross-community collaboration in the teams emerged as an interesting<br />
issue. Relevant data were collected also from project customers, <strong>and</strong> the results <strong>of</strong> the<br />
analysis were presented in P1.<br />
In the paper, SE student projects are considered as a case <strong>of</strong> cross-community<br />
collaboration. Project artifacts negotiated between the team <strong>and</strong> other stakeholders can<br />
37
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
be seen as boundary objects. This perspective helps identify challenges <strong>and</strong> <strong>learning</strong><br />
potential associated with student teams‟ interaction with the different stakeholders, <strong>and</strong><br />
what support might be adequate to help the student teams in this interaction.<br />
<strong>The</strong> paper ends with the recommendation that in the organization <strong>of</strong> the course <strong>and</strong><br />
supervision <strong>of</strong> the projects one should put explicit focus on the role <strong>and</strong> purpose <strong>of</strong><br />
project artifacts serving a role as boundary objects. Consciousness <strong>of</strong> the purpose <strong>of</strong><br />
such artifacts helps shed light on stakeholders‟ different objectives for project<br />
participation <strong>and</strong> the potential role <strong>of</strong> the artifacts in supporting these objectives – two<br />
issues that are, as shown in P1, frequently unclear to the students. As an example,<br />
students typically regard some project artifacts as „school‟ deliverables <strong>and</strong> some as<br />
„real‟ project artifacts relevant to the customer. Documentation <strong>of</strong> the product, for<br />
example, conceptual models describing possible use scenarios, are frequently taken by<br />
the students as belonging to the first category, but may in fact be useful when the<br />
customer wants to continue developing the product. An open discussion between team<br />
<strong>and</strong> customer about such issues may benefit collaboration <strong>and</strong> the project result. <strong>The</strong><br />
current use <strong>of</strong> a number <strong>of</strong> different collaboration tools in the projects point to a<br />
potential for providing shared access among the collaborating stakeholders to boundary<br />
objects <strong>and</strong> descriptions <strong>of</strong> the role <strong>and</strong> design <strong>of</strong> these objects<br />
It is further proposed in P1 that if the students are fully aware <strong>of</strong> the purpose <strong>and</strong> role <strong>of</strong><br />
an artifact in collaboration, they may be allowed to determine for themselves (in<br />
collaboration with the other stakeholder) what is the appropriate form <strong>and</strong> content <strong>of</strong> the<br />
artefact. This contrasts with the m<strong>and</strong>atory use <strong>of</strong> pre-specified templates for, for<br />
example, product requirements or status reports.<br />
5.2 P2: Power through brokering. OSS participation in SE<br />
projects<br />
Author: Krogstie, Birgit<br />
Published in: Proceedings <strong>of</strong> the International Conference on S<strong>of</strong>tware Engineering<br />
(ICSE) 2008<br />
Many s<strong>of</strong>tware engineering projects use open source s<strong>of</strong>tware tools or components. <strong>The</strong> project team’s<br />
active participation in the open source community may be necessary for the team to use the technology.<br />
Based on an in-depth field study <strong>of</strong> industry s<strong>of</strong>tware engineering project students interacting with an<br />
open source community, we find that participation in the community may affect the team’s <strong>work</strong> <strong>and</strong><br />
<strong>learning</strong> by strengthening the power <strong>of</strong> the broker between the team <strong>and</strong> the community. We outline<br />
pitfalls <strong>and</strong> benefits <strong>of</strong> having student teams acquire development-related knowledge from open source<br />
communities. <strong>The</strong> findings are relevant to the organization <strong>and</strong> supervision <strong>of</strong> s<strong>of</strong>tware engineering<br />
student projects interacting with open source communities.<br />
38
Results<br />
<strong>The</strong> background <strong>of</strong> P2 is a longitudinal study conducted with Team A in 2007. <strong>The</strong><br />
team‟s interaction with an open source s<strong>of</strong>tware (OSS) developer community was<br />
essential to the project‟s success. <strong>The</strong> paper gives a chronological account <strong>of</strong> the<br />
project, focusing on the team‟s need for knowledge about an OSS development<br />
frame<strong>work</strong> they were required to use <strong>and</strong> how they approached the challenge <strong>of</strong><br />
acquiring this knowledge. This happened via confusion <strong>and</strong> hesitation through direct<br />
contact with the developer community in the web forum <strong>and</strong> subsequent participation in<br />
that community. <strong>The</strong> participation helped the team get the necessary resources for their<br />
development <strong>and</strong> succeed with their project. <strong>The</strong> team‟s requests for functionality in the<br />
development frame<strong>work</strong> needed for their project led the OSS community to make<br />
changes to the development frame<strong>work</strong> itself.<br />
<strong>The</strong> findings point to some pitfalls <strong>and</strong> benefits for SE education <strong>of</strong> having students<br />
participate in OSS communities. Being a realistic aspect <strong>of</strong> modern SE <strong>work</strong>, OSS<br />
community participation in a capstone project provides <strong>work</strong> experience highly relevant<br />
to industry. <strong>The</strong> team member having the role as broker between the team <strong>and</strong> the<br />
community is likely to be motivated by his/her role as an important contributor to the<br />
team‟s <strong>work</strong> <strong>and</strong> possibly also as a contributor to the open source community.<br />
Motivation <strong>and</strong> pride may be shared by the entire team, as seen in P2. On the other<br />
h<strong>and</strong>, there are some pitfalls in projects depending on direct interaction with an OSS<br />
community. Student teams may hesitate to try <strong>and</strong> participate actively, being in doubt<br />
about their own skills <strong>and</strong> credibility, <strong>and</strong> this may cause problematic project delays.<br />
Attempts to interact with the OSS community may fail if the team does not manage to<br />
convince the community that they qualify as real users <strong>and</strong> potential contributors. <strong>The</strong><br />
OSS community becomes an extra project stakeholder to which the project has to relate,<br />
which makes project management more challenging. For instance, participating on the<br />
OSS arena <strong>and</strong> <strong>work</strong>ing on technical issues there – <strong>and</strong> receiving gratification for the<br />
effort from other participants there – may move the focus from project tasks in need <strong>of</strong><br />
attention. Finally, brokering may be performed inadequately, for example if knowledge<br />
is not properly shared within the team or the team‟s interests are not properly taken care<br />
<strong>of</strong> in the OSS community.<br />
Overall, P2 shows how cross-community collaboration in student projects extends the<br />
traditional customer <strong>and</strong> course staff interaction, with implications for the dynamics <strong>and</strong><br />
<strong>learning</strong> <strong>of</strong> the team. In particular, successful interaction with an OSS community in an<br />
Internet forum requires a certain communications competence.<br />
39
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
5.3 P3: Do’s <strong>and</strong> Don’ts <strong>of</strong> Instant Messaging in Students’<br />
Project Work<br />
Author: Krogstie, Birgit<br />
Published in: Proceedings <strong>of</strong> NOKOBIT 2009.<br />
Instant messaging is a type <strong>of</strong> lightweight collaboration technology heavily used among students in<br />
higher education for social interaction <strong>and</strong> for school <strong>work</strong>. This paper sheds light on students’ use <strong>of</strong><br />
instant messaging to support collaboration in s<strong>of</strong>tware engineering student projects. We draw on several<br />
years’ <strong>of</strong> research on the use <strong>of</strong> collaboration technology in such projects <strong>and</strong> present illustrative<br />
examples <strong>of</strong> how instant messaging is used within the teams <strong>and</strong> for communication with other<br />
stakeholders, e.g. project supervisor <strong>and</strong> customer. We find that the use <strong>of</strong> instant messaging in the<br />
projects is generally successful, but in some cases it is inadequate, which may severely harm the project<br />
outcome. Based on our findings, we discuss implications for the organization <strong>and</strong> supervision <strong>of</strong> SE<br />
student projects..<br />
<strong>The</strong> background <strong>of</strong> P3 was the finding that instant messaging (IM) is extensively used<br />
in the SE student projects. Data on the use <strong>of</strong> collaboration tools by teams collected in<br />
semi-structured interviews <strong>of</strong> project teams in 2006-2008 indicated that there are clear<br />
patterns <strong>of</strong> use with this type <strong>of</strong> tool in the teams. <strong>The</strong>se patterns are in line with the use<br />
<strong>of</strong> IM in <strong>work</strong> life <strong>and</strong> in young people‟s lives as reported in research literature: IM is<br />
used for informal communication, frequently as a substitute for face-to-face<br />
conversation. My role as course staff <strong>and</strong> researcher provided me with knowledge <strong>of</strong><br />
particular IM-related issues in specific project teams in the 2006-2008 cohorts. Given<br />
access to IM logs, I collected examples <strong>of</strong> successful as well as unsuccessful use <strong>of</strong> the<br />
tool. Excerpts from such logs form the core <strong>of</strong> the paper. P3 concludes that generally,<br />
IM is appropriate for informal, within-team collaboration <strong>and</strong> should be applied with<br />
great caution (i.e. only under certain conditions) in formal stakeholder collaboration.<br />
This has implications for the advice course staff should give to student teams about<br />
collaboration.<br />
5.4 P4: <strong>The</strong> wiki as an integrative tool in project <strong>work</strong><br />
Author: Krogstie, Birgit.<br />
Published in: Proceedings <strong>of</strong> COOP 2008.<br />
<strong>The</strong> paper provides insights on how wikis support project <strong>work</strong> <strong>and</strong> what characteristics <strong>of</strong> wikis make<br />
them adequate for this purpose. <strong>The</strong> findings are based on a case study <strong>of</strong> s<strong>of</strong>tware engineering projects<br />
in an educational setting. Project wikis are found to serve an integrative role along several dimensions <strong>of</strong><br />
project <strong>work</strong>, enabled by the flexibility <strong>and</strong> automatic support for capturing history <strong>of</strong>fered by the<br />
technology. <strong>The</strong> findings demonstrate that a project wiki can serve as a knowledge repository, a means<br />
for staging the project, a coordination mechanism, <strong>and</strong> a shared <strong>work</strong>space. To many projects in need <strong>of</strong><br />
40
Results<br />
project management <strong>and</strong> collaboration support, project wikis should be seen as an attractive, lightweight,<br />
all-purpose alternative.<br />
<strong>The</strong> background <strong>of</strong> P4 was the use <strong>of</strong> project wikis in many project teams in the 2007<br />
cohort <strong>of</strong> the SE project course. <strong>The</strong> wiki seemed to be popular as a collaboration tool<br />
<strong>and</strong> used for many purposes in the projects, <strong>and</strong> this led to my decision to investigate<br />
the project wikis in more detail.<br />
Figure 10: Screen shot from a project wiki main page as the project approached<br />
final deadline. Note the strikethroughs <strong>of</strong> completed tasks in the to-do-list.<br />
In P4 it is shown how lightweight project management tools can serve the needs <strong>of</strong><br />
small-scale SE project <strong>work</strong>. In particular, project wikis have a flexibility which makes<br />
it possible to adapt them to a team‟s emerging needs, effectively providing integration<br />
<strong>of</strong> <strong>work</strong> across several dimensions. It was found that the project wikis served to<br />
integrate social <strong>and</strong> task-oriented aspects <strong>of</strong> <strong>work</strong> (see Figure 10, which includes<br />
examples <strong>of</strong> both), integrate team-internal <strong>and</strong> team-external information, <strong>and</strong> integrate<br />
<strong>work</strong> across artifacts. In this way, the wiki played different roles within the project: as a<br />
knowledge repository for the team <strong>and</strong> other stakeholders, as a stage for the team to<br />
41
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
provide a certain image <strong>of</strong> their project (internally <strong>and</strong> externally), as a coordination<br />
mechanism, <strong>and</strong> as shared <strong>work</strong>space.<br />
5.5 P5: Using Project Wiki History to Reflect on the Project<br />
Process<br />
Author: Krogstie, Birgit<br />
Published in: Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System<br />
Sciences (HICSS) 2009<br />
In this paper, we address the potential <strong>of</strong> project wikis to support <strong>reflection</strong> on the project process<br />
through participants’ reconstruction <strong>of</strong> the project trajectory. Drawing on a case study <strong>of</strong> s<strong>of</strong>tware<br />
engineering project <strong>work</strong>, we illustrate how information on project history can be found in a project wiki<br />
<strong>and</strong> may be used to support recall <strong>of</strong> <strong>and</strong> <strong>reflection</strong> on the process. We discuss implications for<br />
postmortem project reviews by considering how the utilization <strong>of</strong> project wikis could addresses some<br />
challenges to successful reviews. We also propose extended wiki functionality that can be used to make a<br />
more selective review <strong>of</strong> project history based on user-tagged contents.<br />
<strong>The</strong> research presented in P5 was initiated based on findings from the research on dayto-day<br />
use <strong>of</strong> project wikis in the capstone projects. In our interviews <strong>of</strong> project team<br />
members about their use <strong>of</strong> project wikis, the wikis themselves were examined as an aid<br />
to the interview. It turned out that recall <strong>and</strong> <strong>reflection</strong> was effectively aided by the<br />
browsing <strong>of</strong> wiki contents. As a consequence I decided to focus not only on day-to-day<br />
use <strong>of</strong> the wikis in the projects, but investigate the potential <strong>of</strong> the wikis to be used<br />
retrospectively to aid recall <strong>of</strong> <strong>and</strong> <strong>reflection</strong> on the process.<br />
P5 demonstrates that given the current use <strong>of</strong> wikis as lightweight project management<br />
tools, there is a potential to utilize the historical information in the wikis for purposes <strong>of</strong><br />
recalling <strong>and</strong> reflecting on important aspects <strong>of</strong> the project process. Wiki contents (e.g.<br />
the various revisions <strong>of</strong> the project wiki main page) can be chronologically traversed in<br />
a „tour‟ <strong>of</strong> the project. By examining page contents <strong>and</strong> following hyper-links, the user<br />
can inspect details related to the process <strong>and</strong> some <strong>of</strong> the project artifacts. <strong>The</strong><br />
combination <strong>of</strong> chronological traversal <strong>and</strong> inspection <strong>of</strong> details aids recall <strong>and</strong><br />
<strong>reflection</strong>.<br />
P5 also identifies a potential for improvement <strong>of</strong> project wikis with the aim <strong>of</strong> providing<br />
better support for retrospective <strong>reflection</strong>. More focused navigation <strong>of</strong> historical data<br />
would be possible if the wiki allowed chronological browsing <strong>of</strong> the wiki pages<br />
considered most relevant to retrospective <strong>reflection</strong>. What is relevant in this respect will<br />
vary between projects <strong>and</strong> individual team members. Importantly, it should be possible<br />
to chronologically browse relevant page revisions (e.g. <strong>of</strong> the main page, showing the<br />
development <strong>of</strong> the team‟s frequently updated to do-list). <strong>The</strong> solution to this challenge<br />
42
Results<br />
proposed in P5 was to allow users to bookmark wiki page revisions with tags that are<br />
individual or shared in the group. Tagging can be done as an integral part <strong>of</strong> day-to-day<br />
<strong>work</strong> activity, for example tagging current page revisions, or retrospectively tagging<br />
older page revisions based on current insights on what might be worth revisiting.<br />
<strong>The</strong> wiki walkthrough tool (WWT) was developed, incorporating the above described<br />
features. <strong>The</strong> WWT is a prototype plug-in to an existing wiki tool. It was integrated <strong>and</strong><br />
tested in the project wiki <strong>of</strong> the team developing the tool in the last phase <strong>of</strong> their<br />
project. Various use scenarios <strong>of</strong> the tool were subject to expert evaluation (groups <strong>of</strong><br />
supervisors <strong>and</strong> customers). Some <strong>of</strong> the scenarios outlined use <strong>of</strong> the tool in project<br />
management <strong>and</strong> some described pedagogical use (e.g. for project supervisors collecting<br />
relevant information about a team‟s <strong>work</strong> throughout their project). <strong>The</strong> evaluation<br />
indicated less enthusiasm for the project management potential <strong>of</strong> the tool <strong>and</strong> more for<br />
its pedagogical potential. Also, it was suggested that the tool be extended to allow not<br />
only bookmarking but also annotation <strong>of</strong> page revisions, e.g. to briefly explain what is<br />
interesting about the pages. Figure 11 shows a screen shot from the tool. In the dialogue<br />
window in the centre <strong>of</strong> the screen, the user is choosing to tag the present version <strong>of</strong> the<br />
wiki page using the tag „gjennomgang‟ (in English: „examination‟ or „revision‟), which<br />
is chosen from a list <strong>of</strong> recently used tags. In the window the user can also specify that<br />
the tag is individual to that user (in this case „arnemart‟) <strong>and</strong> accordingly not shared<br />
with, or visible to, the rest <strong>of</strong> the team.<br />
Figure 11: Wiki Walkthrough Tool: adding a tag to the current version <strong>of</strong> a<br />
page<br />
43
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
5.6 P6: Shared timeline <strong>and</strong> individual experience: Supporting<br />
retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering<br />
teams<br />
Authors: Krogstie, Birgit <strong>and</strong> Divitini, Monica<br />
I was the first author <strong>of</strong> this paper <strong>and</strong> main responsible for setting up the study <strong>and</strong><br />
collecting <strong>and</strong> analyzing the data. <strong>The</strong> second author has provided feedback <strong>and</strong><br />
comments throughout the process.<br />
Published in: Proceedings <strong>of</strong> the 22 nd Conference on S<strong>of</strong>tware Engineering Education<br />
<strong>and</strong> Training (CSEE&T) 2009<br />
To help SE student teams learn from their project process, we propose the use <strong>of</strong> facilitated postmortem<br />
<strong>work</strong>shops in which each team reconstructs its project timeline. Individual team members’ experience <strong>of</strong><br />
the project is included by team members drawing individual ‘experience curves’ along the timeline. <strong>The</strong><br />
approach is based on current industry practice <strong>and</strong> adapted in accordance with theory on <strong>reflection</strong> <strong>and</strong><br />
<strong>learning</strong>. We present the detailed design <strong>of</strong> the <strong>work</strong>shops as they were implemented in an<br />
undergraduate SE project course as well as some recommendations for future <strong>work</strong>shops. <strong>The</strong><br />
<strong>work</strong>shops are effective for promoting active participation, constructing a shared representation <strong>of</strong> the<br />
project process, <strong>and</strong> revisiting project issues from multiple perspectives. Evaluation showed that students<br />
were satisfied with the <strong>work</strong>shop <strong>and</strong> motivated to use the approach in later projects.<br />
<strong>The</strong> background <strong>of</strong> the paper was the assumption that existing post-mortem analysis<br />
approaches from SE industry may be successfully taken into use in SE student projects,<br />
with some adaptation. This would give the students experience with a particular aspect<br />
<strong>of</strong> SE practice <strong>and</strong> help them learn from their project experience. <strong>The</strong> feasibility <strong>of</strong> the<br />
approach in terms <strong>of</strong> time (for students <strong>and</strong> staff) was a key constraint in the adaptation<br />
to the educational setting.<br />
<strong>The</strong> paper presents a concrete method for supporting retrospective <strong>reflection</strong> in student<br />
SE teams. <strong>The</strong> method is adapted from state-<strong>of</strong>-the-art SE industry practice <strong>and</strong> is based<br />
on the individual <strong>and</strong> collective construction <strong>of</strong> a timeline <strong>of</strong> project events <strong>and</strong> the<br />
drawing <strong>of</strong> individual experience curves along the timeline, indicating how positive or<br />
negative the experience was at different points in time. <strong>The</strong> timeline <strong>and</strong> curves serve as<br />
a resource for the team‟s discussion <strong>of</strong> lessons learned. In the student projects, the<br />
approach was used in a short 60 minute <strong>work</strong>shop at the end <strong>of</strong> the project. <strong>The</strong>re was<br />
no formal deliverable from the <strong>work</strong>shop, but the students were recommended to draw<br />
on what they had learned when writing their <strong>reflection</strong> notes (one for each team) to be<br />
h<strong>and</strong>ed in a few days later. <strong>The</strong> timetable <strong>of</strong> the <strong>work</strong>shop is shown in Table 3.<br />
44
Results<br />
Analysis <strong>of</strong> the results showed that the <strong>work</strong>shop helped students share their individual<br />
perspectives (see Figure 12) <strong>and</strong> develop new, shared insights, i.e. knowledge, about<br />
their projects.<br />
<strong>The</strong> techniques applied in the <strong>work</strong>shop were chosen from existing approaches in SE<br />
industry. Thus, the novelty <strong>of</strong> the design primarily lies in the integration <strong>of</strong> the<br />
technique into an educational setting, requiring only minor adaptation <strong>of</strong> the <strong>work</strong>shop<br />
design itself (e.g. aiming for a short duration <strong>and</strong> integrating questions relevant for the<br />
<strong>reflection</strong> notes).<br />
Task<br />
name<br />
Intro<br />
(5 min)<br />
Individual<br />
timelines<br />
(5 min)<br />
Shared<br />
timeline<br />
(15 min)<br />
Individual<br />
experience<br />
curves<br />
(5 min)<br />
Present<br />
curves<br />
(10 min)<br />
Questions<br />
about roles<br />
& lessons<br />
learned<br />
(5 min)<br />
Present<br />
answers<br />
(10 min)<br />
Wrap-up<br />
(5 min)<br />
Description<br />
Explain the purpose <strong>and</strong> agenda <strong>of</strong> the <strong>work</strong>shop. Clarify issues <strong>of</strong><br />
confidentiality <strong>and</strong> research<br />
Each participant gets an A3 sheet <strong>of</strong> paper with a timeline reporting<br />
common events in the course (mainly the deliverables). <strong>The</strong> participants<br />
are asked to individually add events that they perceived had an impact on<br />
their project.<br />
Participants take turn in explaining the events they have listed <strong>The</strong><br />
facilitator marks the events on the whiteboard on a timeline similar to the<br />
one on the individual sheets.<br />
<strong>The</strong> team members each draw their experience curve (or „satisfaction<br />
curve‟) on the A3 sheet. <strong>The</strong> smiley face on top <strong>of</strong> the sheet indicates a<br />
level <strong>of</strong> great satisfaction. Down at the bottom is great dissatisfaction, <strong>and</strong><br />
the timeline itself marks a neutral position in the middle.<br />
Each member in turn goes to the whiteboard, which holds the shared<br />
timeline. <strong>The</strong> team member first draws her curve with her whiteboard<br />
marker, next explains its shape. At the end <strong>of</strong> the session, all team<br />
members‟ experience curves can be found on the whiteboard.<br />
A sheet <strong>of</strong> incomplete statements addressing the project are uncovered,<br />
<strong>and</strong> the students are asked to turn their A3 sheet <strong>and</strong> write their answers<br />
on the blank page.<br />
1) “In the project, my role was…” 2) “Through the project, I got better<br />
at…” 3) “In a similar project, I would like to become more skilled at…”<br />
4) “<strong>The</strong> most important thing I have learnt about s<strong>of</strong>tware engineering in<br />
this project is…”<br />
<strong>The</strong> answers to the questions are presented around the table<br />
<strong>The</strong> students may add further comments about their project on their sheet.<br />
If formal evaluation <strong>of</strong> the <strong>work</strong>shop is not done on another occasion, this<br />
is an opportunity to have feedback, oral or written. <strong>The</strong> A3 sheets are left<br />
for the facilitator/course staff.<br />
Table 3: Timetable <strong>of</strong> the retrospective <strong>reflection</strong> <strong>work</strong>shop conducted with<br />
each <strong>of</strong> the teams in a SE project course. From P6.<br />
45
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Figure 12: Snapshot <strong>of</strong> the whiteboard after the <strong>reflection</strong> <strong>work</strong>shop <strong>of</strong> one<br />
project team, showing the shared project timeline with the individual satisfaction<br />
curves. <strong>The</strong> numbers along Max’s curve indicate where he explained this<br />
experience to the rest <strong>of</strong> the team. From P6.<br />
5.7 P7: A model <strong>of</strong> retrospective <strong>reflection</strong> in project based<br />
<strong>learning</strong> utilizing historical data in collaborative tools<br />
Author: Krogstie, Birgit<br />
Published in: Proceedings <strong>of</strong> EC-TEL 2009<br />
In project based <strong>learning</strong>, <strong>learning</strong> from experience is vital <strong>and</strong> necessitates <strong>reflection</strong>. Retrospective<br />
<strong>reflection</strong> is a conscious, collaborative effort to systematically re-examine a process in order to learn<br />
from it. In s<strong>of</strong>tware development student projects it has been empirically shown that project teams’<br />
retrospective <strong>reflection</strong> can help the teams collaboratively construct new knowledge about their process<br />
<strong>and</strong> that historical data in collaborative tools used in daily project <strong>work</strong> can aid the teams’ recall <strong>and</strong><br />
<strong>reflection</strong> on the different aspects <strong>of</strong> project <strong>work</strong>. In this paper, we draw on these results as well as other<br />
findings on the use <strong>of</strong> collaborative tools in a similar setting. We use the frame<strong>work</strong> <strong>of</strong> distributed<br />
cognition to develop a model <strong>of</strong> retrospective <strong>reflection</strong> in which collaborative tools used as cognitive<br />
tools for daily project <strong>work</strong> are utilized as cognitive tools in retrospective <strong>reflection</strong>, aiding the creation<br />
<strong>of</strong> individual <strong>and</strong> shared representations <strong>of</strong> the project process.<br />
P7 sheds light on the potential to use historical data in various types <strong>of</strong> collaboration<br />
tools in retrospective <strong>reflection</strong>. A reason to utilize different tools is that they are<br />
applied for various purposes in day-to-day project <strong>work</strong> <strong>and</strong>, as a consequence, contain<br />
historical data on different aspects <strong>of</strong> the project process. <strong>The</strong> relevance <strong>of</strong> historical<br />
information in project wikis <strong>and</strong> issue trackers was demonstrated in P5 <strong>and</strong> P8. Email<br />
46
Results<br />
archives are found to contain data that may shed light on the team‟s collaboration with<br />
project stakeholders, which <strong>of</strong>ten entails challenges. <strong>The</strong> contents <strong>of</strong> email archives are<br />
generally considered as formal documentation <strong>of</strong> the project, <strong>and</strong> it is reasonably easy to<br />
search <strong>and</strong> retrieve contents related to some topic or stakeholder relationship. Some<br />
tools are less adequate for use in retrospective <strong>reflection</strong>, partially due to lack <strong>of</strong><br />
features for systematic <strong>and</strong> efficient retrieval <strong>of</strong> the data <strong>and</strong> partially because they are<br />
considered informal. For instance, the case studies <strong>of</strong> the thesis show that the instant<br />
messaging logs <strong>of</strong> members <strong>of</strong> a project team may contain information with a potential<br />
to shed light on important issues in the project (P3, P7). However, instant messaging in<br />
the context <strong>of</strong> project <strong>work</strong> is considered a substitute for informal face-to-face<br />
interaction <strong>and</strong> frequently interwoven with communication related to other activity<br />
(school <strong>and</strong> private). <strong>The</strong> resulting IM logs are not intended or expected to be reexamined<br />
by people other than their owners.<br />
Figure 13: Retrospective <strong>reflection</strong> as distributed cognition. From P7.<br />
On this background, P7 presents a model <strong>of</strong> retrospective <strong>reflection</strong>, taking into account<br />
that different collaboration tools may contain historical data potentially shedding light<br />
on different aspects, or sub-trajectories, <strong>of</strong> the project process. <strong>The</strong> model is based on<br />
47
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
the theoretical frame<strong>work</strong> <strong>of</strong> DCog <strong>and</strong> the model <strong>of</strong> the reflective process proposed by<br />
Boud et al. (1985a) <strong>and</strong> highlights the construction <strong>and</strong> transformation <strong>of</strong><br />
representations as a core mechanism in retrospective <strong>reflection</strong> on a project process.<br />
Also, the model is inspired by the approach to retrospective <strong>reflection</strong> applied in student<br />
teams in the studies described in P6 <strong>and</strong> P8. <strong>The</strong> model explicitly includes steps <strong>of</strong><br />
individual <strong>and</strong> collective <strong>reflection</strong> as well as resulting representations <strong>of</strong> the project<br />
process. Essential to the model, although not heavily explored in the empirical <strong>work</strong> <strong>of</strong><br />
the thesis, is the feedback loop in which lessons learned from the project (in the form <strong>of</strong><br />
various representations <strong>of</strong> the project process) are fed back into the project <strong>work</strong><br />
practice, which could be in the same project or in another one<br />
Another important point made in the paper is that the dual role <strong>of</strong> a collaboration tool<br />
(in day-to-day <strong>work</strong> <strong>and</strong> retrospective <strong>reflection</strong>) might impact on both aspects <strong>of</strong> the<br />
practice through participants‟ awareness that historical data will be re-examined at a<br />
later stage. This awareness might lead to better information sharing as well as to<br />
information hoarding.<br />
5.8 P8: Supporting Reflection in S<strong>of</strong>tware Development with<br />
Everyday Working Tools<br />
Authors: Krogstie, Birgit <strong>and</strong> Divitini, Monica<br />
I was the first author <strong>of</strong> this paper <strong>and</strong> main responsible for setting up the study <strong>and</strong><br />
collecting <strong>and</strong> analyzing the data. <strong>The</strong> second author has provided feedback <strong>and</strong><br />
comments throughout the process.<br />
To appear in: Proceedings <strong>of</strong> COOP 2010<br />
Through their day-to-day usage collaboration tools collect data on the <strong>work</strong> process. <strong>The</strong>se data can be<br />
used to aid participants’ retrospective <strong>reflection</strong> on the process. <strong>The</strong> paper shows how this can be done in<br />
s<strong>of</strong>tware development project <strong>work</strong>. Through a case study we demonstrate how retrospective <strong>reflection</strong><br />
was conducted by use <strong>of</strong> an industry approach to project retrospectives combined with the examination<br />
<strong>of</strong> historical data in Trac, an issue tracker. <strong>The</strong> data helped the team reconstruct the project trajectory<br />
by aiding the recall <strong>of</strong> significant events, leading to a shift in the team’s perspective on the project. <strong>The</strong><br />
success <strong>of</strong> the tool-aided retrospective <strong>reflection</strong> is attributed to its organization as well as the type <strong>of</strong><br />
historical data examined through the tool <strong>and</strong> the tool features for navigating the data. <strong>The</strong>se insights<br />
can be used to help project teams determine the potential <strong>of</strong> their tools to aid retrospective <strong>reflection</strong>.<br />
<strong>The</strong> <strong>work</strong> presented in this paper took as its starting point the previous findings (P5)<br />
about the potential usefulness <strong>of</strong> historical data in wikis to aid retrospective <strong>reflection</strong>.<br />
In the spring semester <strong>of</strong> 2008, project wikis were no longer the popular choice for<br />
project management in the teams <strong>of</strong> our project course. Instead, many student teams<br />
used Trac – an issue tracking tool built on top <strong>of</strong> the file versioning system Subversion<br />
48
Results<br />
(SVN). Trac provides lightweight project management functionality closely integrated<br />
with the s<strong>of</strong>tware development <strong>work</strong>, <strong>and</strong> fits well for projects with a strong focus on<br />
the latter. For this reason, the course staff (including the author) decided to <strong>of</strong>fer an<br />
infrastructure with SVN <strong>and</strong> Trac to the 2008 project teams, accompanied by a brief<br />
introductory lecture. Several teams choose to use Trac. <strong>The</strong> research presented in P8<br />
results from a case study <strong>of</strong> one <strong>of</strong> these teams: team F.<br />
Figure 14: Examination <strong>of</strong> historical data in Trac. From P8.<br />
After following team F in a longitudinal study throughout the semester, we conducted a<br />
three day <strong>work</strong>shop at the end <strong>of</strong> the semester to investigate the possibility <strong>of</strong> improving<br />
the SD retrospective technique <strong>of</strong> constructing individual <strong>and</strong> shared timelines <strong>and</strong><br />
drawing experience curves (used in the other projects in the course as described in P6).<br />
<strong>The</strong> <strong>work</strong>shop design was exp<strong>and</strong>ed to incorporate steps <strong>of</strong> exploring historical data in<br />
49
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Trac, <strong>and</strong> to incorporate individual sessions preceding the collective one. In the<br />
individual sessions, after the students had constructed their timeline <strong>of</strong> the project, we<br />
asked them to chronologically browse the timeline on the Trac main page see to see if<br />
this would make them remember events that had not been remembered by memory<br />
alone. This step would make it possible for us to observe differences between team<br />
members in their use <strong>of</strong>, <strong>and</strong> benefit from, the historical information; differences that<br />
might be attributed to their different roles in the project.<br />
Findings from the study showed that the historical data accessed through the tool helped<br />
the students remember events or detailed information about events (e.g. their exact<br />
time). Whereas all team members remembered something new to add to the purely<br />
memory-based timeline, the benefit <strong>of</strong> using Trac retrospectively was bigger for the two<br />
lead programmers than for the other three team members. <strong>The</strong> event „pre-study coding‟<br />
(e.g. the onset <strong>of</strong> coding <strong>work</strong>) was not remembered by anyone prior to use <strong>of</strong> Trac. <strong>The</strong><br />
use <strong>of</strong> Trac led the two lead programmers to remember <strong>and</strong> acknowledge that coding<br />
<strong>work</strong> took place early in the project. In the collective timeline construction, pre-study<br />
coding was brought into the shared timeline <strong>and</strong> achieved „<strong>of</strong>ficial status‟ as important<br />
in the team. However, as the follow-up interviews from the study demonstrated, there<br />
were viewpoints among the three other team members on how the pre-study activity<br />
was conducted that were not brought up, probably due to the dominance <strong>of</strong> the two lead<br />
programmers <strong>and</strong> their version <strong>of</strong> what pre-study programming was about. This is<br />
further elaborated in the evaluation <strong>of</strong> the research process in 7.3.1.<br />
<strong>The</strong> findings from the study are summarized in three main points about what was<br />
important to make the tool-aided <strong>reflection</strong> successful: the organization <strong>of</strong> the <strong>work</strong>shop<br />
into individual <strong>and</strong> collaborative steps <strong>of</strong> knowledge construction with the use <strong>of</strong><br />
adequate representations <strong>of</strong> the project process; the tool features for giving<br />
chronological overview <strong>of</strong> project events; <strong>and</strong> the tool features for navigating between<br />
overview <strong>and</strong> detail. <strong>The</strong> latter features permitted access from the chronological<br />
overview to there-<strong>and</strong>-then versions <strong>of</strong> project artifacts, thus supporting the close<br />
linking <strong>of</strong> process <strong>and</strong> product that is characteristic <strong>of</strong> SE <strong>work</strong>. Based on the findings,<br />
P8 outlines a set <strong>of</strong> guiding questions that can be used to help assess the potential <strong>of</strong> a<br />
collaboration tool to aid <strong>reflection</strong> on SE <strong>work</strong>. <strong>The</strong> questions are intended as a step<br />
towards a frame<strong>work</strong> relating various types <strong>of</strong> collaboration tools to the aspects <strong>of</strong> dayto-day<br />
SE <strong>work</strong> supported by the tools <strong>and</strong> the potential for the tools to aid retrospective<br />
<strong>reflection</strong>.<br />
P8 also provides an example <strong>of</strong> how the <strong>reflection</strong> model from P7 can be instantiated so<br />
as to comprise selected elements <strong>of</strong> a specific <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong>, in this<br />
case focusing on the role <strong>of</strong> Trac in the case study <strong>of</strong> P8 (see Figure 15).<br />
50
Results<br />
Figure 15: A model outlining how Trac supports project <strong>work</strong> <strong>and</strong> retrospective<br />
<strong>reflection</strong> on the project process in the case study <strong>of</strong> P8. From P8.<br />
51
6 Contributions <strong>and</strong> implications<br />
This chapter summarizes the research contributions <strong>and</strong> their implications. <strong>The</strong> chapter<br />
is structured in accordance with the four main research topics shown in the diagram in<br />
Figure 3.<br />
6.1 Project <strong>work</strong> as cross-community collaboration<br />
Contribution 1 <strong>of</strong> the thesis is an increased underst<strong>and</strong>ing <strong>of</strong> cross-community<br />
collaboration in SE student teams, conceptualized in terms <strong>of</strong> communities <strong>of</strong> practice<br />
<strong>and</strong> related to the use <strong>of</strong> collaboration tools. <strong>The</strong> contribution is an extension <strong>of</strong> state-<strong>of</strong>the-art<br />
research on how stakeholder communication <strong>and</strong> collaboration, as a crucial part<br />
<strong>of</strong> SE practice, is currently approached in SE education, <strong>and</strong> how it can be more<br />
carefully attended to in capstone projects.<br />
In order for student teams to conduct their cross-community collaboration in a reflective<br />
<strong>and</strong> successful way they need to relate <strong>and</strong> align to the stakeholders‟ practices <strong>and</strong> the<br />
associated objectives for involvement with the student project. As students‟<br />
underst<strong>and</strong>ing <strong>of</strong> these objectives may be limited (P1), project courses should be<br />
designed to increase students‟ consciousness about this issue <strong>and</strong> encourage openness<br />
among project stakeholders about their objectives in the collaboration processes. Also,<br />
team members are likely to have different simultaneous goals for their project<br />
participation, linked to their multiple memberships in different communities. Making<br />
these objectives explicit within the relevant collaboration settings basically corresponds<br />
53
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
to properly clarifying expectations for the collaborative <strong>work</strong> activity. In the thesis, the<br />
variation in team members‟ objectives for their project <strong>work</strong> is also illustrated in the<br />
<strong>reflection</strong> <strong>work</strong>shops described in P6. <strong>The</strong> individual satisfaction curves along with the<br />
explanations provided by their creators illustrate how the „ups <strong>and</strong> downs‟ <strong>of</strong> project<br />
<strong>work</strong> relate to the team members‟ goals <strong>and</strong> expectations <strong>and</strong> also those <strong>of</strong> other<br />
stakeholders (e.g. when the curves <strong>of</strong> all team members drop at a point where negative<br />
feedback was provided by the supervisor, or when a team member explains that she was<br />
unhappy with her task over a long period because she would have liked to learn new<br />
technology in the project rather than just applying a technology she knew in advance).<br />
Apart from being a source <strong>of</strong> insight for the project teams about such issues in their<br />
projects, data from the <strong>reflection</strong> <strong>work</strong>shops <strong>of</strong> all the projects in a course can be used<br />
by course staff or researchers to gain insights about the effect <strong>of</strong> the current course<br />
design, e.g. in terms <strong>of</strong> stakeholder collaboration in the projects.<br />
Participants in SE student teams generally have a clear conception <strong>of</strong> the usage <strong>of</strong><br />
different collaboration tools in their project. <strong>The</strong> choice <strong>of</strong> tool for a particular instance<br />
<strong>of</strong> collaboration appears to depend on whether the collaboration is within the team or<br />
between team <strong>and</strong> stakeholders, <strong>and</strong> whether it is formal or informal. A similar<br />
categorization can be used by the researcher (e.g. project course staff) investigating the<br />
usage <strong>of</strong> particular tools in project teams with the ultimate objective <strong>of</strong> providing<br />
guidance about the appropriateness <strong>of</strong> particular tools for different purposes. Exploring<br />
the use <strong>of</strong> instant messaging in the student teams (P3), it is shown that team-stakeholder<br />
collaboration by use <strong>of</strong> instant messaging poses particular challenges. More generally, it<br />
is suggested that the appropriateness <strong>of</strong> a tool for team-stakeholder collaboration be a<br />
question <strong>of</strong> negotiation between the stakeholders. Even with agreement between<br />
stakeholders, using a tool in a setting for which it is not generally considered<br />
appropriate (e.g. instant messaging for formal meetings) should be done with great<br />
caution.<br />
By looking at cross-community collaboration as achieved through brokering (P2), the<br />
thesis provides an underst<strong>and</strong>ing <strong>of</strong> how the dynamics <strong>of</strong> cross-community<br />
collaboration is closely related to the role <strong>and</strong> competence <strong>of</strong> the individual team<br />
member. Not only the team as a community, but the individual‟s role in the team <strong>and</strong><br />
competence in adapting to the practice <strong>of</strong> both communities must be considered to<br />
underst<strong>and</strong> how cross-community collaboration takes place through brokering <strong>and</strong> how<br />
it might be supported.<br />
Contribution 1 can be a resource for practitioners <strong>and</strong> researchers within SE Education,<br />
providing rationale <strong>and</strong> guidelines for developing project courses in the direction <strong>of</strong><br />
54
Contributions <strong>and</strong> implications<br />
more consciousness, insight <strong>and</strong> <strong>reflection</strong> among students <strong>and</strong> course staff about<br />
stakeholder objectives <strong>and</strong> tool use in cross-community collaboration.<br />
6.2 Lightweight collaboration tools in project <strong>work</strong><br />
Contribution 2 <strong>of</strong> this thesis is an increased underst<strong>and</strong>ing <strong>of</strong> the use <strong>of</strong> lightweight<br />
collaboration tools in the <strong>work</strong> practice <strong>of</strong> SE student teams. <strong>The</strong> contribution is a<br />
response to the continuous need within CSCW <strong>and</strong> TEL for new knowledge about the<br />
use <strong>of</strong> state-<strong>of</strong>-the-art technologies in current practices. In the case <strong>of</strong> SE student<br />
projects, this is a need boosted by the spread <strong>of</strong> lightweight technologies in many areas<br />
<strong>of</strong> modern life including project <strong>work</strong>, combined with the changes in educational <strong>and</strong><br />
SE industry practices. In the thesis, key issues related to tool usage in student projects<br />
have been framed by considering the teams as communities <strong>of</strong> practice <strong>and</strong> <strong>learning</strong> <strong>and</strong><br />
by considering the <strong>work</strong> <strong>and</strong> <strong>learning</strong> in the teams as a case <strong>of</strong> distributed cognition.<br />
<strong>The</strong> thesis provides specific insights on the use <strong>of</strong> instant messaging tools, internet<br />
forums, project wikis <strong>and</strong> issue trackers, <strong>and</strong> also on the use <strong>of</strong> email as part <strong>of</strong> a<br />
concerted use <strong>of</strong> tools in the project <strong>work</strong>. Through various case studies the thesis<br />
shows how the current use <strong>of</strong> the tools is related to core challenges <strong>of</strong> SE projects <strong>and</strong><br />
project based <strong>learning</strong> (e.g. cross-community collaboration <strong>and</strong> the <strong>learning</strong> <strong>of</strong> new<br />
technologies). Regarding issue trackers, the thesis primarily addresses their use from the<br />
perspective <strong>of</strong> retrospective <strong>reflection</strong>, but in elaborating on the connection between<br />
retrospective <strong>and</strong> day-to-day use <strong>of</strong> the tools (P7 <strong>and</strong> P8), the thesis also sheds light on<br />
the role <strong>of</strong> issue trackers in day-to-day project <strong>work</strong>.<br />
Research paper P6 does not provide insights on computer-based lightweight tools but<br />
can be considered an illustration <strong>of</strong> the successful use <strong>of</strong> a different type <strong>of</strong> lightweight<br />
tools in project teams‟ <strong>work</strong>: physical tools in the form <strong>of</strong> papers, pens <strong>and</strong> whiteboards.<br />
<strong>The</strong> focus <strong>of</strong> the thesis is on the process in which the physical tools are used <strong>and</strong> the<br />
representations that are created (P6-P8). <strong>The</strong> physical tools continue to play an<br />
55
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
important role in the version <strong>of</strong> the <strong>reflection</strong> <strong>work</strong>shop which is aided by historical data<br />
in a computerized collaboration tool (P7-P8).<br />
For practitioners within SE education <strong>and</strong> PBL more generally, knowledge about the<br />
typical <strong>and</strong> recommended use <strong>of</strong> collaboration tools in student projects is an aid to<br />
underst<strong>and</strong>ing the challenges <strong>of</strong> specific projects <strong>and</strong> project courses. This<br />
underst<strong>and</strong>ing is essential for course staff seeking to provide project students with the<br />
appropriate scaffolding for <strong>learning</strong>. While recognizing the skills with which students<br />
currently h<strong>and</strong>le a variety <strong>of</strong> collaboration tools for <strong>work</strong> <strong>and</strong> social purposes, course<br />
staff can make some recommendations <strong>and</strong> help the students be more conscious <strong>of</strong> their<br />
tool use in different collaborative settings. <strong>The</strong> research on lightweight collaboration<br />
tool use in student projects presented in the thesis can also serve as a basis for the<br />
identification <strong>of</strong> issues for educational practitioners‟ own pedagogical research <strong>and</strong><br />
development <strong>of</strong> practice (e.g. addressing the role <strong>of</strong> a particular collaboration tool in<br />
student projects or how a particular aspect <strong>of</strong> project <strong>work</strong> is supported by the use <strong>of</strong><br />
collaboration tools). <strong>The</strong> research methods as well as the results presented in the thesis<br />
can inform this type <strong>of</strong> research.<br />
For the organization <strong>of</strong> small scale SE project <strong>work</strong> in educational <strong>and</strong> other settings,<br />
the findings on how a project wiki may support project management (P5) can aid the<br />
decision <strong>of</strong> whether, <strong>and</strong> how, to use a project wiki as a lightweight project management<br />
tool in a specific project. <strong>The</strong>se findings can also be useful to project teams who are<br />
already using project wikis <strong>and</strong> who may benefit from utilizing the tool in different<br />
ways. <strong>The</strong> thesis research on issue trackers (P7 <strong>and</strong> P8) can inform considerations on<br />
the choice <strong>of</strong> lightweight project management tool for a project <strong>and</strong> the comparison <strong>of</strong><br />
project wikis <strong>and</strong> issue trackers based on their qualities for day-to-day <strong>work</strong> <strong>and</strong><br />
retrospective <strong>reflection</strong>.<br />
<strong>The</strong> thesis contribution on the use <strong>of</strong> project wikis <strong>and</strong> their potential integrative role in<br />
small-scale projects adds to the knowledge within CSCW on tool support for project<br />
<strong>work</strong> as well as the usage <strong>of</strong> wiki technology. For organizers <strong>of</strong> small-scale projects<br />
the contribution can inform decisions about whether <strong>and</strong> how to make use <strong>of</strong> wiki<br />
technology for lightweight project management.<br />
For the TEL research field, the contribution adds to the body <strong>of</strong> literature on the use <strong>of</strong><br />
collaboration tools (e.g. wikis) in educational settings, <strong>of</strong>fering a perspective that views<br />
the tools primarily as aids to daily <strong>work</strong> practice (i.e. as tools for <strong>work</strong> rather than<br />
<strong>learning</strong> technology) <strong>and</strong> indirectly supporting <strong>learning</strong> in the context <strong>of</strong> PBL.<br />
Further, for the research communities in which the research papers have been published,<br />
the thesis provides concrete examples <strong>of</strong> how daily <strong>work</strong> <strong>and</strong> tool use in a team can be<br />
56
Contributions <strong>and</strong> implications<br />
investigated retrospectively by the researcher. Specifically, the retrospective use <strong>of</strong><br />
project wikis (P5) <strong>and</strong> issue trackers (P7-P8) exemplifies how the researcher can gain<br />
insights about daily <strong>work</strong> in a team with the aid <strong>of</strong> historical data contextualized<br />
through participants‟ retrospective accounts <strong>of</strong> the process.<br />
6.3 Supporting retrospective <strong>reflection</strong><br />
Contribution 3 <strong>of</strong> the thesis comprises new knowledge about how retrospective<br />
<strong>reflection</strong>, seen as a part <strong>of</strong> the collaborative <strong>work</strong> practice, can be supported in SE<br />
student projects <strong>and</strong> project <strong>work</strong> more generally. It also includes tools for supporting<br />
such <strong>reflection</strong>.<br />
A concrete <strong>and</strong> empirically tested design for retrospective <strong>reflection</strong> <strong>work</strong>shops, adapted<br />
from current SE industry practice to the educational setting <strong>of</strong> SE student projects, is<br />
<strong>of</strong>fered (P6).<br />
A new <strong>and</strong> dual use <strong>of</strong> existing collaboration tools in SE projects is proposed, extending<br />
the current role <strong>of</strong> the tools in daily <strong>work</strong> to that <strong>of</strong> supporting participants‟ <strong>reflection</strong> on<br />
the project process. Post-mortem interviews (P5) <strong>and</strong> retrospective <strong>work</strong>shops (P8) are<br />
shown to be possible ways <strong>of</strong> integrating participants‟ examination <strong>of</strong> historical data<br />
into a systematic reconstruction <strong>of</strong> the project process with the aim <strong>of</strong> identifying<br />
lessons learned. For purposes <strong>of</strong> examining historical data, relevant tool features have<br />
been identified along with important elements in the organization <strong>of</strong> the retrospective<br />
<strong>reflection</strong> effort. Finally, the thesis <strong>of</strong>fers a set <strong>of</strong> guiding questions for assessing the<br />
potential <strong>of</strong> specific collaboration tools to aid retrospective <strong>reflection</strong> in project <strong>work</strong><br />
(P8).<br />
In the case <strong>of</strong> project wikis, the PhD <strong>work</strong> contributes with a tool (the Wiki<br />
Walkthrough tool) extending current wiki functionality to address identified<br />
shortcomings <strong>of</strong> wikis with respect to their use in retrospective <strong>reflection</strong> (P5). <strong>The</strong> tool<br />
57
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
has been developed as a prototype <strong>and</strong> tested through scenario evaluation by expert<br />
groups (P5).<br />
For course staff <strong>and</strong> others involved in the organization <strong>of</strong> SE student projects, the tools<br />
for conducting retrospective <strong>reflection</strong> provide a way <strong>of</strong> implementing state-<strong>of</strong>-the-art<br />
SE approaches (e.g. agile development methodologies) in SE student projects as well as<br />
a way <strong>of</strong> improving the <strong>learning</strong> outcomes. <strong>The</strong> insights about the potential <strong>of</strong><br />
collaboration tools to support retrospective <strong>reflection</strong> can help course staff decide<br />
whether to impose or encourage the use <strong>of</strong> particular tools in the projects, based on their<br />
potential to aid retrospective <strong>reflection</strong>.<br />
For organizers <strong>of</strong> PBL in domains other than s<strong>of</strong>tware engineering, the thesis<br />
demonstrates that an important part <strong>of</strong> the reflective activity <strong>of</strong> a project team can be<br />
facilitated <strong>and</strong> organized in a <strong>work</strong>shop format. <strong>The</strong> design <strong>of</strong> the <strong>work</strong>shops as outlined<br />
in P6 is focused on project <strong>work</strong> <strong>and</strong> project process independently <strong>of</strong> the SE domain<br />
<strong>and</strong> is applicable to other <strong>work</strong> domains. As argued in the thesis, the ideal in PBL is that<br />
the retrospective <strong>reflection</strong> <strong>of</strong> a team be part <strong>of</strong> the real <strong>work</strong> practice <strong>and</strong> not a separate,<br />
educational exercise. Thus, for PBL-based courses within other domains the organizers<br />
should find ways <strong>of</strong> making the <strong>reflection</strong> <strong>work</strong>shops an integral part <strong>of</strong> the <strong>work</strong><br />
practice, possibly by linking them to existing <strong>learning</strong> practices within the domain. As a<br />
second option, the thesis research indicates that <strong>reflection</strong> <strong>work</strong>shops can be introduced<br />
as more <strong>of</strong> an educational exercise: students appear to find it interesting to get the<br />
opportunity to reconstruct <strong>and</strong> account for their story about the project <strong>and</strong> hear the<br />
story <strong>of</strong> the other team members. <strong>The</strong> approach based on the drawing <strong>of</strong> individual <strong>and</strong><br />
shared timelines <strong>and</strong> satisfaction curves allows participants to address the emotional<br />
aspects <strong>of</strong> project <strong>work</strong> in a form that does not put them <strong>of</strong>f writing or talking „about<br />
their feelings‟ or spending what they see as too much time elaborating on the project<br />
process. <strong>The</strong> <strong>reflection</strong> <strong>work</strong>shop presented in P8 was adapted to a CSCW research<br />
agenda <strong>and</strong> is not a readily applicable design for retrospective <strong>reflection</strong> in PBL<br />
settings. However, the finding that steps <strong>of</strong> investigating historical data in collaboration<br />
tools can be incorporated into an existing <strong>work</strong>shop approach <strong>and</strong> thereby improve it,<br />
can be utilized by PBL organizers as in the development <strong>of</strong> other tool-aided <strong>reflection</strong><br />
<strong>work</strong>shop designs. For the organizations (in industry or education) in which PBL takes<br />
place, the proposed approaches to retrospective <strong>reflection</strong>s may contribute to<br />
organizational <strong>learning</strong>. Organizers <strong>of</strong> PBL efforts consider using the results from the<br />
<strong>reflection</strong> <strong>work</strong>shops as a way <strong>of</strong> informing the development <strong>of</strong> their courses, taking<br />
some precautions as discussed in P6. This would support the educational practitioners as<br />
researchers <strong>and</strong> developers <strong>of</strong> their own practice. Also, in formal education, successful<br />
incorporation <strong>of</strong> retrospective <strong>reflection</strong> <strong>work</strong>shops in PBL can ultimately contribute to<br />
industry practice if the students who learn the techniques find them useful <strong>and</strong> later<br />
58
Contributions <strong>and</strong> implications<br />
apply them in their pr<strong>of</strong>essional career. This is another way <strong>of</strong> bringing about crosscommunity<br />
<strong>learning</strong> through PBL.<br />
To the CSCW field, the thesis provides new knowledge about how data stored in<br />
collaboration tools through a <strong>work</strong> practice, can be utilized to support that <strong>work</strong> practice<br />
through retrospective <strong>reflection</strong>. Framed in state-<strong>of</strong>-the-art industry practice for project<br />
retrospectives in SE project <strong>work</strong>, current usage <strong>of</strong> lightweight collaboration tools in<br />
project <strong>work</strong> (e.g. as manifest in Contribution 2), theory on human <strong>reflection</strong> <strong>and</strong><br />
<strong>learning</strong>, <strong>and</strong> a focus on the project process (e.g. trajectory <strong>and</strong> sub-trajectories) as the<br />
main object <strong>of</strong> <strong>reflection</strong>, some key points are established. Firstly, day-to-day usage <strong>of</strong><br />
collaboration tools in a project determines what aspects (<strong>and</strong> challenges) <strong>of</strong> project <strong>work</strong><br />
become reflected in the tools. Secondly, particular tool features are useful for access to<br />
<strong>and</strong> navigation <strong>of</strong> historical data stored in (or by) the tools in the context <strong>of</strong><br />
retrospective <strong>reflection</strong>. This includes chronological overview <strong>of</strong> the project process <strong>and</strong><br />
easy access from points in the process to project artifacts (e.g. the project product) in<br />
their there-<strong>and</strong>-then state. Thirdly, the organization <strong>of</strong> the retrospective <strong>reflection</strong><br />
activity in which the historical data are used, should be designed to incorporate<br />
individual as well as collaborative knowledge construction <strong>and</strong> attend to „facts‟ as well<br />
as „feelings‟ about the project process. Fourthly, if collaboration tools are given a dual<br />
role in a <strong>work</strong> practice, supporting day-to-day <strong>work</strong> as well as retrospective <strong>reflection</strong><br />
on that <strong>work</strong>, the effect on the <strong>work</strong> practice should be considered <strong>and</strong> preferably<br />
empirically investigated (e.g. looking for signs <strong>of</strong> increased or improved information<br />
sharing or information hoarding).<br />
For tool designers, these key points can be used to inform the design <strong>of</strong> collaboration<br />
tools with the aim <strong>of</strong> making the tools useful for retrospective <strong>reflection</strong>.<br />
For researchers analyzing project <strong>work</strong> settings with the aim <strong>of</strong> identifying ways <strong>of</strong><br />
supporting the <strong>work</strong> practice, the key points illuminate the option <strong>of</strong> introducing or<br />
redesigning retrospective <strong>reflection</strong> activities, e.g. by utilizing collaboration tools as a<br />
resource for the <strong>reflection</strong>. In the field <strong>of</strong> S<strong>of</strong>tware Engineering in particular, the<br />
proposed use <strong>of</strong> collaboration tools can be considered an answer to one <strong>of</strong> the<br />
challenges to having SE pr<strong>of</strong>essionals actually prioritize retrospective <strong>reflection</strong> on their<br />
projects: getting access to relevant data to aid the <strong>reflection</strong> process.<br />
59
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
6.4 A model <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>reflection</strong><br />
Contribution 4 <strong>of</strong> the thesis is the <strong>reflection</strong> model shown in Figure 13, conceptualized<br />
in terms <strong>of</strong> distributed cognition <strong>and</strong> stressing the cognitive as well as the social aspects<br />
<strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong> in project teams.<br />
<strong>The</strong> model is presented in P7 as a contribution to the TEL research field. <strong>The</strong> model‟s<br />
perspective on the <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> is novel in making explicit the<br />
potential role <strong>of</strong> collaboration tools in supporting day-to-day <strong>work</strong> <strong>and</strong> the retrospective<br />
<strong>reflection</strong> on that <strong>work</strong> <strong>and</strong> thereby strengthening the integration <strong>of</strong> these aspects <strong>of</strong> the<br />
<strong>work</strong> practice. <strong>The</strong> model further incorporates individual <strong>and</strong> collaborative elements <strong>of</strong><br />
the reflective process, <strong>and</strong> makes explicit the use <strong>of</strong> representations <strong>of</strong> the project<br />
process (ranging from the trajectories „in participants‟ heads‟ to explicit representations<br />
developed through dedicated <strong>reflection</strong> activities) as means for, <strong>and</strong> outcomes <strong>of</strong>,<br />
<strong>reflection</strong>. <strong>The</strong> model represents a practical view <strong>of</strong> the <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong><br />
<strong>and</strong> its support in project based <strong>learning</strong>, incorporating elements from <strong>learning</strong> theory<br />
<strong>and</strong> SE industry practices that have been combined <strong>and</strong> empirically tested in SE student<br />
projects as part <strong>of</strong> the thesis research. While the <strong>work</strong>shops in the thesis research were<br />
based on a SE retrospective approach using timelines <strong>and</strong> satisfaction curves as explicit<br />
representations <strong>of</strong> the project process, the <strong>reflection</strong> model does not presume any<br />
particular form <strong>of</strong> explicit representation <strong>of</strong> the project.<br />
<strong>The</strong> model can be used to aid design for retrospective <strong>reflection</strong> in PBL settings. <strong>The</strong><br />
designers could be educational practitioners seeking to develop their own practice,<br />
project managers, people responsible for <strong>learning</strong> activities on a project or organization<br />
level, <strong>and</strong> designers <strong>of</strong> collaboration tools.<br />
In the analysis <strong>of</strong> a PBL setting, the elements <strong>of</strong> the model <strong>and</strong> their interrelationships<br />
can be used to identify a set <strong>of</strong> considerations about the <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong><br />
<strong>and</strong> trigger relevant questions. Some such questions are: When should individuals‟<br />
<strong>reflection</strong> on their <strong>work</strong> practice be explicitly encouraged or designed for, <strong>and</strong> what is<br />
60
Contributions <strong>and</strong> implications<br />
the desired outcome <strong>of</strong> this <strong>reflection</strong>? Similarly, when should collaborative <strong>reflection</strong><br />
efforts be encouraged <strong>and</strong>/or arranged, in what form <strong>and</strong> with what type <strong>of</strong> outcomes?<br />
What is the intended use <strong>of</strong> the outcomes the <strong>reflection</strong>? For instance, how should<br />
individual <strong>and</strong> collaborative <strong>reflection</strong> inform each other, <strong>and</strong> how should the outcomes<br />
inform daily <strong>work</strong>? What collaborative tools currently in use are potential sources <strong>of</strong><br />
historical data that may enrich retrospective <strong>reflection</strong>? Should other tools be taken into<br />
use in day-to-day <strong>work</strong> practice in order to gather data for use in <strong>reflection</strong> on <strong>work</strong>?<br />
How do we know that the outcome <strong>of</strong> <strong>reflection</strong> is in fact useful, to the project <strong>and</strong>/or to<br />
the organization? <strong>The</strong> number <strong>of</strong> possible questions is large. <strong>The</strong> <strong>work</strong> to address such<br />
questions to aid analysis <strong>of</strong>, <strong>and</strong> design for, <strong>reflection</strong> in PBL should be seen as an<br />
integral part <strong>of</strong> <strong>learning</strong>-related activities in the organization (e.g. under the heading <strong>of</strong><br />
course development, organizational <strong>learning</strong> or knowledge management). <strong>The</strong> <strong>reflection</strong><br />
model can be instantiated to outline the elements <strong>of</strong> a specific <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong><br />
<strong>cycle</strong>, as shown in P8 (Figure 15) where the historical data in an issue tracker were used<br />
in a retrospective <strong>reflection</strong> <strong>work</strong>shop.<br />
It may add to the practical usefulness <strong>of</strong> the <strong>reflection</strong> model to view it in light <strong>of</strong><br />
Contribution 3. This makes it easier to see how the ideas in the model are grounded in<br />
practice <strong>and</strong> can be implemented in a concrete design for <strong>reflection</strong> in project <strong>work</strong>.<br />
For TEL <strong>and</strong> CSCW researchers aiming to develop theory on <strong>reflection</strong>, the model can<br />
serve as a basis for empirical or theoretical <strong>work</strong>.<br />
Finally, the <strong>reflection</strong> model can serve as an aid to analysis <strong>of</strong> cases <strong>of</strong> <strong>work</strong> <strong>and</strong><br />
<strong>learning</strong>, helping the researcher determine how to gather data about the <strong>work</strong> process.<br />
<strong>The</strong> historical data in collaboration tools contextualized through participants‟<br />
reconstruction <strong>of</strong> the project process can be a resource for participants (team members<br />
or other stakeholders (e.g. project supervisors) acting as researchers on their own<br />
practice, for example through action research.<br />
61
7 Evaluation<br />
In this chapter, I evaluate the contributions <strong>of</strong> the thesis with respect to the research<br />
questions <strong>and</strong> the state-<strong>of</strong>-the-art <strong>of</strong> the research fields. Also, I evaluate the research<br />
approach <strong>and</strong> process <strong>and</strong> discuss my choice <strong>of</strong> theory.<br />
7.1 Evaluation <strong>of</strong> the contributions with respect to the research<br />
questions<br />
RQ1: What characterizes s<strong>of</strong>tware engineering student projects? <strong>The</strong> thesis answers<br />
this question by providing new knowledge about two selected aspects <strong>of</strong> project <strong>work</strong> in<br />
SE student projects: issues related to cross-community collaboration (Contribution 1)<br />
<strong>and</strong> the use <strong>of</strong> lightweight collaboration tools (Contribution 2).<br />
RQ2: What is the current usage <strong>of</strong> lightweight collaboration tools to support <strong>work</strong> in<br />
s<strong>of</strong>tware engineering student projects? This question is answered by Contribution 2,<br />
which represents new knowledge on the use <strong>of</strong> a selection <strong>of</strong> commonly used<br />
collaboration tools in SE student projects: instant messaging, internet forums, project<br />
wikis, issue trackers, <strong>and</strong> also email as part <strong>of</strong> a concerted use <strong>of</strong> tools in the teams.<br />
RQ3: How can retrospective <strong>reflection</strong> be supported in s<strong>of</strong>tware engineering student<br />
projects? <strong>The</strong> thesis answers the question in Contribution 3 by showing that a SE<br />
industry approach to organizing <strong>reflection</strong> <strong>work</strong>shops can be successfully adapted to<br />
student projects, providing adequate support for the reflective process as outlined in<br />
relevant <strong>learning</strong> theory. A concrete <strong>work</strong>shop design is outlined. Also, the thesis shows<br />
that collaboration tools <strong>and</strong> their historical data can be used as a resource for the<br />
<strong>reflection</strong>.<br />
RQ4: How can collaboration tools that are used in daily project <strong>work</strong> be utilized as<br />
tools for project based <strong>learning</strong>? This is answered in Contribution 3, in which it is<br />
proposed to have collaboration tools take on a dual role in the projects: as support for<br />
day-to-day <strong>work</strong> <strong>and</strong> as support for retrospective <strong>reflection</strong> on that <strong>work</strong> by use <strong>of</strong> the<br />
historical data in the tools. It is shown that certain tool features are important in this<br />
63
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
respect, <strong>and</strong> that an existing retrospective <strong>reflection</strong> <strong>work</strong>shop approach can be used to<br />
frame the exploration <strong>of</strong> historical data in the collaboration tools. <strong>The</strong> thesis does not<br />
<strong>of</strong>fer a ready-to-use <strong>work</strong>shop template in this respect, but <strong>of</strong>fers guidelines for the<br />
evaluation <strong>of</strong> the potential <strong>of</strong> tools to be used in retrospective <strong>reflection</strong> <strong>and</strong> a<br />
recommendation to integrate the exploration <strong>of</strong> historical data into a timeline <strong>and</strong><br />
satisfaction curve <strong>work</strong>shop approach. In Contribution 4 the answer to RQ4 is framed<br />
theoretically <strong>and</strong> generalized to project based <strong>learning</strong> at large through the model<br />
outlining the <strong>work</strong>-<strong>learning</strong>-<strong>reflection</strong> <strong>cycle</strong>.<br />
7.2 Evaluation <strong>of</strong> the contributions with respect to state-<strong>of</strong>-theart<br />
<strong>of</strong> the SE Education, TEL <strong>and</strong> CSCW research fields<br />
This section provides an overview <strong>of</strong> how the research contributions add to the literature<br />
outlined in the state-<strong>of</strong>-the-art chapter (Chapter 3).<br />
Contribution 1 adds to SE education <strong>and</strong> TEL research on the challenges <strong>of</strong> project<br />
based <strong>learning</strong> (Section 3.1). <strong>The</strong> findings on how students address stakeholder<br />
objectives <strong>and</strong> how we can help the students be more conscious about these objectives<br />
(P1) contributes to the literature on stakeholder collaboration in SE project courses. <strong>The</strong><br />
study in P2 adds to the limited but growing literature on student participation in OSS<br />
development. Additionally, the study in P2 provides a contribution to the TEL body <strong>of</strong><br />
literature on project based <strong>learning</strong> by applying the CoP concept <strong>of</strong> brokering to shed<br />
light on the relationship between individual <strong>work</strong> <strong>and</strong> <strong>learning</strong> <strong>and</strong> that <strong>of</strong> the team.<br />
Finally, the thesis research on instant messaging represents a small increment to state<strong>of</strong>-the-art<br />
TEL research on the use <strong>of</strong> collaboration tools in project based <strong>learning</strong> <strong>and</strong> to<br />
the literature on instant messaging in educational contexts.<br />
Contribution 2 brings forward state-<strong>of</strong>-the-art research on the use <strong>of</strong> lightweight<br />
collaboration technology to support project based <strong>learning</strong> (Section 3.2). <strong>The</strong> thesis<br />
research on project wikis (P4) in day-to-day project <strong>work</strong> also adds to the CSCW<br />
literature on tool support for small-scale project <strong>work</strong>, in which the use <strong>of</strong> project wikis<br />
has so far not been extensively addressed. <strong>The</strong> research on wikis can also be seen as a<br />
contribution to the TEL literature, in which previous studies <strong>of</strong> wikis in educational<br />
contexts have mainly been addressing wikis as <strong>learning</strong> technology <strong>and</strong> not – as in this<br />
thesis – as tools for collaborative <strong>work</strong>.<br />
Contribution 3 brings forward state-<strong>of</strong>-the-art research in SE education <strong>and</strong> TEL on<br />
how to support project based <strong>learning</strong> (Section 3.3), focusing on how to support one<br />
particular aspect <strong>of</strong> the <strong>work</strong> practice: retrospective <strong>reflection</strong>. Also, Contribution 3<br />
64
Evaluation<br />
adds to the CSCW literature <strong>of</strong> how collaboration tools in daily use in <strong>work</strong> practice can<br />
be utilized to support that practice (Section 3.4).<br />
With respect to the S<strong>of</strong>tware Engineering literature, the contribution meets two<br />
recognized challenges to successful project retrospectives: the problem <strong>of</strong> inadequate<br />
data (Kasi et al. 2008) <strong>and</strong> the problem <strong>of</strong> project participants finding it ineffective<br />
<strong>and</strong>/or not personally useful to code their experience (Schindler <strong>and</strong> Eppler 2003) – the<br />
solution <strong>of</strong> this thesis being to utilize existing historical data.<br />
<strong>The</strong> study by Ar<strong>and</strong>a <strong>and</strong> Veolia (2009) examining the potential use <strong>of</strong> repository data<br />
from bug fixing processes to aid coordination <strong>of</strong> SE <strong>work</strong> (Section 3.4) complements<br />
findings in the thesis. As the study was recently published <strong>and</strong> accordingly is not<br />
discussed in the research papers <strong>of</strong> the thesis, it will be briefly elaborated here. <strong>The</strong><br />
study confirms that participants‟ meaning making is essential for the reconstruction <strong>of</strong><br />
project trajectories (e.g. stories <strong>of</strong> bug fixing). It is difficult to use repository data as a<br />
source <strong>of</strong> immediately useful input for decision making in day-to-day coordination<br />
because the data are insufficient <strong>and</strong> sometimes erroneous. For instance, repository data<br />
may frequently fail to reflect who were actually involved in the <strong>work</strong> process. In this<br />
PhD thesis, the focus is also on how to design for meaning-making by use <strong>of</strong> historical<br />
data, but in a setting outside the immediate time pressure <strong>of</strong> day-to-day coordination. In<br />
the case <strong>of</strong> retrospective <strong>reflection</strong>, the reliance on human memory makes it less<br />
important that the historical data from collaboration tools alone can provide the basis for<br />
reconstructing complete trajectories. Historical data providing erroneous information<br />
about the project process might <strong>of</strong> course be a problem even if participants have the<br />
time to check it against their memory <strong>and</strong> other data sources. <strong>The</strong> use <strong>of</strong> historical data<br />
from a set <strong>of</strong> different collaboration tools reflecting various aspects <strong>of</strong> the project<br />
process can help participants build sufficiently coherent knowledge about the process A<br />
similarity between the findings <strong>of</strong> Ar<strong>and</strong>a <strong>and</strong> Veolia <strong>and</strong> those underlying Contribution<br />
3 is that both studies confirm the usefulness <strong>of</strong> historical data linking the development<br />
process to the evolving product. Seen together, the studies strengthen the argument that<br />
data in collaboration tools may serve a dual role in project <strong>work</strong> as shown in the<br />
<strong>reflection</strong> model in P7, <strong>and</strong> that to achieve this, the process needed for participants to<br />
actively give meaning to the data must be given due attention, whether in the context <strong>of</strong><br />
retrospective <strong>reflection</strong> or in the day-to-day coordination <strong>of</strong> <strong>work</strong>.<br />
Contribution 4 brings forward state-<strong>of</strong>-the-art CSCW <strong>and</strong> TEL research on how to<br />
support, integrate <strong>and</strong> conceptualize <strong>work</strong>, <strong>reflection</strong> <strong>and</strong> <strong>learning</strong> in collaborative <strong>work</strong>.<br />
<strong>The</strong> <strong>reflection</strong> model in Figure 13 is a theoretical contribution mainly to the TEL field,<br />
building on existing models considering <strong>learning</strong> as a cyclic process in which<br />
experience is followed by <strong>reflection</strong> on the experience (Section 2.2). State-<strong>of</strong>-the-art SE<br />
65
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
practice (Section 3.1), the theoretical frame<strong>work</strong> <strong>of</strong> DCog (Section 2.1.3) <strong>and</strong> research<br />
on the use <strong>of</strong> collaboration tools in collaborative project <strong>work</strong> (Section 3.2), as well as<br />
Contributions 2 <strong>and</strong> 3 from the thesis research, were drawn upon to incorporate<br />
elements specific to retrospective <strong>reflection</strong> <strong>and</strong> its support: the potential dual role <strong>of</strong><br />
collaboration tools combined with the inclusion <strong>of</strong> individual as well as collaborative<br />
steps <strong>of</strong> knowledge construction by aid <strong>of</strong> various representations.<br />
In the fields <strong>of</strong> CSCW <strong>and</strong> CSCL, the mainstream perspective on <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
considers meaning-making to be basically social by nature, <strong>and</strong> the unit <strong>of</strong> analysis is<br />
typically the community or group in which the <strong>work</strong> <strong>and</strong> <strong>learning</strong> takes place. Whereas<br />
some research has a focus on the relationship between the individual <strong>and</strong> the collective<br />
(Engeström 2000; Stahl 2002), there is a need (Kaptelinin <strong>and</strong> Nardi 2006) for further<br />
research emphasizing the interrelationship between the two. <strong>The</strong> research <strong>of</strong> this thesis<br />
can be seen to contribute in this direction, both in the individual case studies (e.g. P2,<br />
P6), <strong>and</strong> in the theoretical synthesis <strong>of</strong> the last papers (P7-P8). <strong>The</strong> use <strong>of</strong> concepts from<br />
DCog enabled a clearer image <strong>of</strong> the collaborative <strong>and</strong> the individual as elements in a<br />
reflective process.<br />
With regard to the dual role <strong>of</strong> collaboration tools, the CSCW <strong>and</strong> SE bodies <strong>of</strong> research<br />
(Section 3.4) on the use <strong>of</strong> data gathered from project <strong>work</strong> for purposes <strong>of</strong> supporting<br />
the day-to-day <strong>work</strong> practice reveal similarities between use <strong>of</strong> historical data to aid<br />
day-to-day <strong>work</strong> (e.g. awareness <strong>and</strong> coordination (Ar<strong>and</strong>a <strong>and</strong> Venolia 2009;<br />
Omoronyia et al. 2009)) <strong>and</strong> the use <strong>of</strong> the data to aid retrospective <strong>reflection</strong>. Both<br />
types <strong>of</strong> settings entail meaning-making based on the reconstruction <strong>of</strong> project<br />
trajectories, with the purpose <strong>of</strong> evaluating <strong>and</strong> acting upon the outcome. On a high<br />
level <strong>of</strong> abstraction, it can be argued that the <strong>reflection</strong> model in P8 covers the gathering<br />
<strong>and</strong> use <strong>of</strong> data from collaboration tools for support <strong>of</strong> any aspect <strong>of</strong> the <strong>work</strong> practice.<br />
In its current form, the <strong>reflection</strong> model is particularly useful for shedding light on<br />
support for retrospective <strong>reflection</strong> <strong>and</strong> how it can be integrated into the <strong>work</strong> practice.<br />
<strong>The</strong> outcome <strong>of</strong> the reflective process (cf. Section 2.2, Figure 5) is intended to be<br />
valuable for the <strong>work</strong> process. Exactly how this should be achieved has not been<br />
thoroughly explored in the empirical research <strong>of</strong> this PhD <strong>work</strong>, but treated more on an<br />
abstract level with reference to the role <strong>of</strong> retrospective <strong>reflection</strong> in SE practice <strong>and</strong> the<br />
assumption that further research is needed on the use <strong>and</strong> usefulness <strong>of</strong> the <strong>reflection</strong><br />
outcomes. Continued research to validate <strong>and</strong> refine the model in order to improve its<br />
value for process improvement <strong>and</strong> organizational <strong>learning</strong> should relate to the SE <strong>and</strong><br />
TEL research literature <strong>and</strong> also draw on state-<strong>of</strong>-the-art research within project<br />
management, organizational <strong>learning</strong> <strong>and</strong> knowledge management.<br />
66
Evaluation<br />
7.3 Evaluation <strong>of</strong> the research process<br />
This PhD <strong>work</strong> is situated on the boundary between several research fields, each with its<br />
literature, approaches <strong>and</strong> research topics. In the case <strong>of</strong> SE education, TEL <strong>and</strong> CSCW,<br />
there is significant overlap <strong>of</strong> research topics related to collaborative <strong>learning</strong>. Studies<br />
<strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong> in SE student projects are a source <strong>of</strong> relevant research material<br />
for conferences such as ICSE, EC-TEL <strong>and</strong> COOP. <strong>The</strong> fact that the research papers <strong>of</strong><br />
this thesis have been published within the different communities indicates a fairly wide<br />
scope <strong>of</strong> the PhD <strong>work</strong>. While an aim to focus on publication within one <strong>of</strong> the research<br />
communities might have resulted in a more extensive contribution within that field,<br />
broader participation opens up for a larger set <strong>of</strong> perspectives on the research topics at<br />
h<strong>and</strong>. For instance, the SE education research community favours an engineering- <strong>and</strong><br />
practitioner-oriented (albeit theoretically grounded) stance to SE student projects,<br />
whereas the CSCW <strong>and</strong> TEL/CSCL fields generally require a stronger theoretical<br />
contextualization. Being a computer scientist <strong>and</strong> a social scientist, I appreciate the<br />
opportunity to approach my research in both ways.<br />
<strong>The</strong> research <strong>of</strong> this thesis has been presented as a combination <strong>of</strong> interpretive <strong>and</strong><br />
design research. This section will address how the studies actually correspond to these<br />
research approaches <strong>and</strong>, as part <strong>of</strong> the discussion, shed light on some methodological<br />
challenges.<br />
7.3.1 On the <strong>work</strong> in the thesis as interpretive research<br />
As outlined in Chapter 4, the case studies <strong>of</strong> the thesis can largely be seen as<br />
interpretive. <strong>The</strong> longitudinal studies are the ones that most clearly fit the<br />
characterization. Klein <strong>and</strong> Myers describe IS research as interpretive “if it is assumed<br />
that our knowledge <strong>of</strong> reality is gained only through social constructions such as<br />
language, consciousness, shared meanings, documents, tools, <strong>and</strong> other artifacts.” It<br />
“does not predefine dependent <strong>and</strong> independent variables, but focuses on the complexity<br />
<strong>of</strong> human sense making as the situation emerges []; it attempts to underst<strong>and</strong><br />
phenomena through the meanings that people assign to them” (Klein <strong>and</strong> Myers 1999,<br />
p.69). This approach to research is in line with a basic constructionist perspective on<br />
<strong>work</strong> <strong>and</strong> <strong>learning</strong>. <strong>The</strong> connection to the IS field does not make interpretive research<br />
invalid for CSCW <strong>and</strong> CSCL, both <strong>of</strong> which are research fields with a focus on people‟s<br />
sense making through their use <strong>of</strong> tools <strong>and</strong> artifacts.<br />
According to Walsham, results from interpretive case studies can be generalized in the<br />
following ways: by developing concepts, by generating theory, by drawing specific<br />
implications, <strong>and</strong> by providing rich insights (Walsham 1995). In the thesis, the<br />
67
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
contribution <strong>of</strong> research paper P2 („Power through brokering‟) can be seen as the<br />
clearest example <strong>of</strong> „rich insights‟. <strong>The</strong> overall focus <strong>of</strong> the thesis research on current<br />
practices in the SE student projects has been on providing insights about these practices<br />
<strong>and</strong> in particular the use <strong>of</strong> collaboration tools. To make the insights more applicable,<br />
they have in several cases been developed into guidelines <strong>and</strong> recommendations. This<br />
seems to be particularly appropriate for contributions to the SE Education community<br />
<strong>and</strong> should be a fair approach as long as the basis for providing the recommendations is<br />
properly taken into account. In the case studies where data have been collected across<br />
many project teams (P1,P3,P4), generalization has to some extent approached that <strong>of</strong><br />
positivist case studies (Yin 2003) by addressing what is more or less typical for a certain<br />
population (e.g SE student projects).<br />
In terms <strong>of</strong> use <strong>of</strong> theory, there are three main approaches in interpretive research<br />
(Walsham 1995): using theory initially to guide design <strong>and</strong> data collection, using it as<br />
part <strong>of</strong> an iterative process <strong>of</strong> data collection <strong>and</strong> analysis, <strong>and</strong> to using it as a final<br />
product <strong>of</strong> the research. A restricted use <strong>of</strong> theory in an early phase resembles the use <strong>of</strong><br />
sensitizing concepts (Bowen 2006) in grounded theory. In this thesis, some theory (cf<br />
Chapters 2-3) has been used from the start <strong>of</strong> empirical studies whereas other theory has<br />
been introduced during data collection <strong>and</strong> analysis. Table 4 shows which main<br />
theoretical concepts have informed the <strong>work</strong> presented in the various research papers.<br />
<strong>The</strong> table indicates which theory was informing data collection <strong>and</strong> analysis (X), <strong>and</strong><br />
which theory was informing the analysis phase only (Xa). <strong>The</strong> choice <strong>and</strong> combination<br />
<strong>of</strong> theory for the thesis is further discussed in 7.4.<br />
Table 4: <strong>The</strong>ory informing the <strong>work</strong> presented in the research papers<br />
68
Evaluation<br />
In the longitudinal studies I have aimed for a balance between the initial use <strong>of</strong> theory<br />
<strong>and</strong> an open attitude (“What‟s the story here”), seeking to build an underst<strong>and</strong>ing <strong>of</strong><br />
participants‟ interpretations <strong>of</strong> the phenomena at h<strong>and</strong> as well as my own interpretation.<br />
I have made no “commitment to indifference” (R<strong>and</strong>all et al. 2007) but have tried to be<br />
aware <strong>of</strong> my own preconceptions <strong>and</strong> ways <strong>of</strong> thinking: those originating in my<br />
pr<strong>of</strong>essional background (e.g. as course staff, as an engineer in computer science, <strong>and</strong> as<br />
a social scientist who <strong>work</strong>ed with activity theory in her Master thesis) <strong>and</strong> those that<br />
are more situational/personal (e.g. developing a particular liking for a team member or<br />
being tired during an observation session).<br />
Klein <strong>and</strong> Myers (1999) present a set <strong>of</strong> seven principles <strong>of</strong> interpretive field studies<br />
that can be used to guide the design <strong>and</strong> evaluation <strong>of</strong> such studies. In what follows, I<br />
use examples from the longitudinal studies (Team A <strong>and</strong> Team F) to illustrate the<br />
relevance <strong>of</strong> the principles as well as how I have followed them.<br />
<strong>The</strong> fundamental principle <strong>of</strong> the hermeneutic circle says that the parts <strong>and</strong> the whole<br />
should be seen in light <strong>of</strong> each other. In the longitudinal research, the writing,<br />
elaboration <strong>and</strong> re-examination <strong>of</strong> field notes helped me move back <strong>and</strong> forth between<br />
details (e.g. a particular utterance in a meeting or an instance <strong>of</strong> collaboration tool use)<br />
<strong>and</strong> the whole picture (e.g. the project process). Another example is that throughout the<br />
PhD research I have aimed to underst<strong>and</strong> the use <strong>of</strong> specific collaboration tools in the<br />
context <strong>of</strong> the project teams‟ concerted use <strong>of</strong> tools.<br />
With respect to the principle <strong>of</strong> contextualization, one challenge in this PhD <strong>work</strong> has<br />
been the question <strong>of</strong> whether the activity in the teams should be considered mainly as<br />
<strong>learning</strong> or as <strong>work</strong> – or as both. One answer is provided by the choice to focus on <strong>work</strong><br />
as the core <strong>of</strong> project-based <strong>learning</strong>, which means the focus <strong>of</strong> empirical studies can be<br />
more related to <strong>work</strong> whereas the connection to <strong>learning</strong> is theoretically underpinned.<br />
Another answer is that the focus on <strong>work</strong> or <strong>learning</strong> depends on the intended audience<br />
for the presentation <strong>of</strong> the research. <strong>The</strong> research <strong>of</strong> the thesis has been contextualized<br />
in slightly different ways to provide relevant contributions to various communities (see<br />
Figure 3). Another issue related to contextualization is that data collection <strong>and</strong> analysis<br />
is contextualized not only by the chosen theory but also by the underlying interests <strong>of</strong><br />
the researcher. <strong>The</strong> latter in particular is not always opaque in the presentation <strong>of</strong> the<br />
research. For instance, my interest in process trajectories <strong>and</strong> retrospective <strong>reflection</strong><br />
reflects a general research interest in <strong>work</strong> processes, which then become the context for<br />
underst<strong>and</strong>ing other aspects <strong>of</strong> the SE student projects. Knowledge about the product<br />
being developed has to me mainly been instrumental in making sense <strong>of</strong> the project<br />
process, even though I believe the completion <strong>of</strong> the product to be a key motivational<br />
factor for the project participants. Finally, a core topic in the thesis research is how the<br />
69
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
students <strong>and</strong> course staff themselves contextualize the SE projects, <strong>and</strong> how this impacts<br />
on <strong>work</strong> <strong>and</strong> <strong>learning</strong> (cf. Section 3.1).<br />
<strong>The</strong> principle <strong>of</strong> interaction between the researchers <strong>and</strong> the subjects does not state that<br />
no interaction should take place, but that the researcher be conscious about the possible<br />
impact <strong>of</strong> the interaction on the research. This impact may be difficult to assess. In the<br />
longitudinal studies, while I was trying to disturb the teams as little as possible in their<br />
day-to-day <strong>work</strong>, I participated socially (e.g. by engaging in chit-chat prior to <strong>work</strong><br />
sessions). Sometimes I brought some chocolate or fruit. While observing <strong>work</strong>, I<br />
occasionally asked clarifying questions, but only in situations where I did not<br />
underst<strong>and</strong> what the team members were doing <strong>and</strong> it seemed important to underst<strong>and</strong> it<br />
to make sense <strong>of</strong> their <strong>work</strong> <strong>and</strong> tool use. In such situations, the team members‟ answers<br />
to my questions turned into meta-comments about their activity <strong>and</strong> were no longer part<br />
<strong>of</strong> the team‟s ordinary activity; they were talking „about their world‟ instead <strong>of</strong> just<br />
being in it (Sacks 1992). <strong>The</strong> distinction between meta comments <strong>and</strong> comments which<br />
were part <strong>of</strong> normal activity could sometimes be blurred: to some extent, commenting<br />
aloud about <strong>work</strong> was part <strong>of</strong> „normal‟ activity <strong>and</strong> provided other team members with<br />
increased awareness <strong>of</strong> the ongoing <strong>work</strong> through consequential communication<br />
(Gutwin <strong>and</strong> Greenberg 2002). This „thinking aloud‟ was particularly visible in<br />
situations where students <strong>work</strong>ed in pairs, one student at the keyboard <strong>and</strong> the other at<br />
his side to assist. On some occasions the thinking aloud might have been intended to<br />
inform me as a researcher. All in all, as I kept following the students, they seemed to<br />
accept my presence but largely focused on their <strong>work</strong> task.<br />
Importantly, my role as a researcher <strong>and</strong> not supervisor or examiner was made clear to<br />
the study teams (as well as to the other teams in the course, who had to be informed that<br />
the research subjects did not get any advantage over other teams). <strong>The</strong>se conditions<br />
were generally unproblematic. On one occasion I chose to leave a <strong>work</strong> session in the<br />
computer lab as the team was in real need <strong>of</strong> help <strong>and</strong> started pushing me – while I<br />
really felt like helping them. On another occasion, I decided that I would provide the<br />
team with a small piece <strong>of</strong> advice to get them going in an ineffective communication<br />
process – a decision rooted solely in my own impatience. This was however an<br />
exception.<br />
<strong>The</strong> principle <strong>of</strong> abstraction <strong>and</strong> generalization specifies that theoretical abstractions<br />
always be related to “field study details as they were experienced”. This connection has<br />
been maintained by looking at concrete instances <strong>of</strong> collaboration tool use in the case<br />
studies <strong>and</strong> placing the observations in context <strong>of</strong> theoretical abstractions (e.g. brokering<br />
(P2) <strong>and</strong> reconstruction <strong>of</strong> a process trajectory (P7, P8)).<br />
70
Evaluation<br />
According to the principle <strong>of</strong> dialogical reasoning it is important to make transparent (to<br />
the reader <strong>and</strong> to oneself as researcher) “the historical intellectual basis <strong>of</strong> the research<br />
(i.e. its fundamental philosophical assumptions)” – in other words, explain “the lenses<br />
through which field data are construed, documented <strong>and</strong> organized” (Klein <strong>and</strong> Myers<br />
1999, p.76). <strong>The</strong> research papers <strong>of</strong> the thesis always include a section outlining<br />
theoretical background framing the empirical studies. Also, while the introduction <strong>of</strong><br />
relevant theory has happened at different stages <strong>of</strong> research for the different papers (see<br />
Table 4), they are all grounded in an underlying constructionist view, which is made<br />
explicit in some <strong>of</strong> the papers.<br />
<strong>The</strong> principle <strong>of</strong> multiple interpretations <strong>and</strong> the principle <strong>of</strong> suspicion are important in<br />
interpretive research, in which the identification <strong>of</strong> interesting phenomena may lead the<br />
researcher to look for confirmatory data <strong>and</strong> findings while being less observant to<br />
contradictory ones. As a case in point I will describe an experience related to the study<br />
<strong>of</strong> Team F from which I gained important insights about the case <strong>and</strong> the research<br />
process, in particular the importance <strong>of</strong> follow-up interviews. This could only be very<br />
briefly addressed in P8. Figure 16 shows two diagrams originating in my analysis <strong>of</strong> the<br />
data from the retrospective <strong>reflection</strong> <strong>work</strong>shop <strong>of</strong> Team F (see P7 <strong>and</strong> P8). <strong>The</strong>y depict<br />
my interpretation <strong>of</strong> how the team changed the story <strong>of</strong> their project during their<br />
<strong>reflection</strong> <strong>work</strong>shop; the left diagram outlining the „before‟ version <strong>and</strong> the right<br />
diagram showing the „after‟-.<br />
Figure 16: Diagrams showing the researchers' interpretation <strong>of</strong> how Team F<br />
changed their story <strong>of</strong> their project in their <strong>reflection</strong> <strong>work</strong>shop.<br />
<strong>The</strong> diagrams were shown to the team members in individual follow-up interviews<br />
about three months after the <strong>work</strong>shop (after the summer holiday). <strong>The</strong> students thought<br />
the diagrams <strong>and</strong> my accompanying explanation fit with their conception <strong>of</strong> what<br />
happened during the <strong>work</strong>shop. <strong>The</strong>se viewpoints could be seen as strengthening my<br />
interpretation, even taking into account the effect <strong>of</strong> the model power (Bråten 1981;<br />
Buhl <strong>and</strong> Richter 2004) exerted by me as a researcher presenting finished diagrams.<br />
However, the informal interviews with the students also brought about some new<br />
71
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
viewpoints on the project process that had not emerged in the <strong>work</strong>shop – at least not<br />
clearly enough to be explicitly addressed in the teams‟ discussion at the time. From the<br />
interviews <strong>and</strong> my subsequent re-examination <strong>of</strong> the <strong>work</strong>shop data, there was reason to<br />
believe that these viewpoints had already been there during the <strong>work</strong>shop (<strong>and</strong> were not<br />
mainly results <strong>of</strong> maturation after the <strong>work</strong>shop), but they had been expressed only halfheartedly<br />
then. <strong>The</strong> viewpoints were those <strong>of</strong> the less dominant team members. My<br />
initial interpretation <strong>of</strong> how the <strong>work</strong>shop had changed the team‟s view <strong>of</strong> the process<br />
reflected the perspective <strong>of</strong> the more dominant team members.<br />
<strong>The</strong>se insights served as an important reminder that I might have been attending more to<br />
the activity <strong>and</strong> viewpoints <strong>of</strong> some team members than others during the many hours <strong>of</strong><br />
observation – maybe because they were the students most clearly influencing the team‟s<br />
development <strong>work</strong> <strong>and</strong> the project process, <strong>and</strong> maybe because they were the ones most<br />
able <strong>and</strong> prone to share their insights <strong>and</strong> viewpoints during day-to-day <strong>work</strong>. Apart<br />
from providing methodological insights, the follow-up interviews made me refine the<br />
findings (in P8) on the use <strong>of</strong> historical data in retrospective <strong>reflection</strong> (acknowledging<br />
that data originating in a certain tool might be more relevant to some team members<br />
than others <strong>and</strong> serve to strengthen already dominant views on the process) <strong>and</strong> on how<br />
a <strong>reflection</strong> <strong>work</strong>shop should be conducted (e.g. by using explicit facilitation to make<br />
everyone‟s voice heard).<br />
Considering all the studies in the PhD research, a triangulation <strong>of</strong> data sources has been<br />
applied throughout. This includes interviews with project participants <strong>and</strong> stakeholders,<br />
contributing to multiple interpretations. Going back to participants to check if the<br />
researcher‟s conceptions <strong>of</strong> the phenomena in question match those <strong>of</strong> the participants<br />
has mainly been done for the longitudinal studies <strong>of</strong> single teams.<br />
7.3.2 On the <strong>work</strong> in the thesis as design research<br />
Design research in education is directed at “developing, testing, implementing, <strong>and</strong><br />
diffusing innovative practices to move the socially constructed forms <strong>of</strong> teaching <strong>and</strong><br />
<strong>learning</strong> from malfunction to function or from function to excellence” (Kelly et al.<br />
2008a).<br />
<strong>The</strong> improvement <strong>of</strong> practice should involve the participants (e.g. students <strong>and</strong> teachers<br />
(Brown 1992; Collins 1992)) in the process. Brown describes her intentions with design<br />
experiments in educational research as to ”transform classrooms from academic <strong>work</strong><br />
factories to <strong>learning</strong> environments that encourage reflective practice among students,<br />
teachers, <strong>and</strong> researchers.” (p.174) <strong>The</strong> focus <strong>of</strong> the research in this thesis on better<br />
equipping teams in SE student teams to reflect on their own practice could be seen as a<br />
way <strong>of</strong> achieving this, although the focus is on <strong>reflection</strong> as part <strong>of</strong> <strong>work</strong> practice.<br />
72
Evaluation<br />
However, only some <strong>of</strong> the <strong>reflection</strong> seen in the teams (e.g. in P6) addresses how the<br />
course is organized; the rest is (as intended) about <strong>work</strong>.<br />
In the thesis, the interpretive research informing the research activities on <strong>reflection</strong> can<br />
be seen as part <strong>of</strong> a design research effort starting with a more general exploratory<br />
agenda <strong>and</strong> refining focus to <strong>reflection</strong> as an area for improvement. Guidelines for the<br />
organization <strong>of</strong> SE student project are proposed (P1-P4). <strong>The</strong> thesis research involving<br />
interventions by trying out new solutions (i.e. the wiki walkthrough tool in P4 <strong>and</strong> the<br />
<strong>reflection</strong> <strong>work</strong>shops with <strong>and</strong> without the aid <strong>of</strong> historical data in collaboration tools in<br />
P6-P8) result in st<strong>and</strong>alone contributions (Contributions 2 <strong>and</strong> 3) but should<br />
simultaneously be considered as first steps along a line <strong>of</strong> research aiming to improve<br />
the solutions. If similar solutions are tried out in new projects within the SE project<br />
course <strong>of</strong> study, there will be an iteration that can result in improvements from one year<br />
to the next, with the researcher-as-course-staff as a change agent. To achieve<br />
improvement <strong>of</strong> the project course within a semester, the <strong>reflection</strong> <strong>work</strong>shops will have<br />
to take place during the semester as well as at the end <strong>of</strong> the semester. Research efforts<br />
along these lines are considered further <strong>work</strong> in the thesis (see Section 8).<br />
Students‟ <strong>reflection</strong> on their project <strong>work</strong> can be a resource for course staff aiming to<br />
improve the course, as proposed in P6. To allow the students to maintain their focus on<br />
the <strong>work</strong> practice, course staff will have to be brokers translating students‟ practicefocused<br />
experience into issues <strong>of</strong> how to organize the course.<br />
Design research should also contribute to the development <strong>of</strong> theory (Cobb et al. 2003,<br />
p.9). In the thesis, development <strong>of</strong> theory on the basis <strong>of</strong> empirical studies <strong>of</strong> <strong>reflection</strong><br />
resulted in the <strong>reflection</strong> model presented in P7.<br />
In sum, presenting the research <strong>of</strong> the thesis as design research is done with some<br />
precautions <strong>and</strong> refers to the long-term research agenda as well as to the present results.<br />
A full-fledged design research effort requires a systematic, iterative development <strong>of</strong> the<br />
proposed solutions within a project course. I believe that the tools <strong>and</strong> theory developed<br />
in this thesis on the basis <strong>of</strong> empirical <strong>work</strong> with projects in the selected SE project<br />
course will be a useful basis for systematic improvement <strong>of</strong> the course <strong>and</strong> the specific<br />
approaches to retrospective <strong>reflection</strong>.<br />
7.4 On the choice <strong>of</strong> theory<br />
So, my first piece <strong>of</strong> advice for new researchers is for them to choose<br />
theories which they feel are insightful to them.<br />
(Walsham 2006)<br />
73
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Across educational, social <strong>and</strong> organizational psychology, there is a multitude <strong>of</strong><br />
frame<strong>work</strong>s describing structures <strong>of</strong> collectively created meaning emerging in, <strong>and</strong><br />
coordinating, groups‟ activities (Akkerman et al. 2007). Two <strong>of</strong> them are used for the<br />
thesis: Communities <strong>of</strong> practice (CoP) (Wenger 1998; Wenger 2000), which is closely<br />
connected to situated <strong>learning</strong> (Lave <strong>and</strong> Wenger 1991), <strong>and</strong> DCog (Hutchins 1995;<br />
Rogers <strong>and</strong> Ellis 1994). Other relevant frame<strong>work</strong>s include actor net<strong>work</strong> theory (ANT)<br />
(Latour 2005) <strong>and</strong> activity theory (AT) (Engeström 1987; Kaptelinin <strong>and</strong> Nardi 2006;<br />
Kuutti 1995; Leont'ev 1981).<br />
<strong>The</strong>re are several studies comparing different theoretical frame<strong>work</strong>s with respect to<br />
their use in CSCW (Bratteteig <strong>and</strong> Gregory 1999; Halverson 2002; Nardi 1992; Nardi<br />
2002; R<strong>and</strong>all et al. 2007). In this thesis, I will not go deeply into the discussions <strong>of</strong> the<br />
pros <strong>and</strong> cons <strong>of</strong> the frame<strong>work</strong>s, but mainly stress the rationale for my choice <strong>of</strong><br />
theory.<br />
My initial view <strong>of</strong> what was interesting in the cases was influenced by my background<br />
in s<strong>of</strong>tware engineering (with its challenges <strong>and</strong> methodologies) <strong>and</strong> educational theory<br />
(having used activity theory to analyze an empirical case in my Master thesis, (Krogstie<br />
2000)). I brought with me a general interest in what happens within a group (or<br />
community) <strong>of</strong> people <strong>work</strong>ing <strong>and</strong> <strong>learning</strong> together <strong>and</strong> at the intersection <strong>of</strong> such<br />
groups (or communities). To describe <strong>work</strong> <strong>and</strong> <strong>learning</strong> from the latter perspective,<br />
CoP concepts can be used, including those <strong>of</strong> boundaries <strong>and</strong> brokering. AT, <strong>and</strong> in<br />
particular the concept <strong>of</strong> activity system (Engeström 1987), are adequate for analysis <strong>of</strong><br />
similar settings, tensions/contradictions within <strong>and</strong> between activity systems serving as<br />
a way <strong>of</strong> accounting for the static as well as dynamic aspects <strong>of</strong> interrelated activity<br />
systems. ANT provides means <strong>of</strong> accounting for the dynamics within <strong>and</strong> between<br />
net<strong>work</strong>s, the concept <strong>of</strong> alignment (Latour 2005) being central. DCog can be used to<br />
shed light on the same mechanisms by focusing on the transformation <strong>of</strong> representations<br />
within <strong>and</strong> between functional systems (Hutchins 1995).<br />
Halverson states that from a pragmatic view, one can identify four desired attributes <strong>of</strong> a<br />
theory: descriptive power, rhetorical power, inferential power, <strong>and</strong> applicability to the<br />
real world, the latter largely translating to the ability to inform design (Halverson 2002).<br />
<strong>The</strong> above mentioned frame<strong>work</strong>s describing structures <strong>of</strong> collectively created meaning,<br />
are all powerful means <strong>of</strong> analyzing <strong>and</strong> describing settings <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong> when<br />
systematically applied. When it comes to the power to inform design, DCog <strong>and</strong> AT<br />
both have their proponents; Halverson (2002) arguing most strongly for the powers <strong>of</strong><br />
DCog <strong>and</strong> Kaptelinin <strong>and</strong> Nardi (2006) arguing in favour <strong>of</strong> AT. Independent <strong>of</strong> the<br />
choice <strong>of</strong> theoretical frame<strong>work</strong>, there is general agreement that the step from analysis<br />
74
Evaluation<br />
to design is problematic even if a theory <strong>of</strong> good descriptive power is used to<br />
systematically outline the as-is situation (Dourish 2006).<br />
In the early phases <strong>of</strong> the empirical <strong>work</strong>, I sought to use theoretical concepts mainly as<br />
sensitizing concepts (Blumer 1954). In choosing theory, I aimed for „conceptual<br />
simplicity‟ to gain the aid <strong>of</strong> some structure without having a frame<strong>work</strong> too strongly<br />
enforcing a preconception <strong>of</strong> what was interesting about the cases. A restricted use <strong>of</strong><br />
theoretical concepts creates an opening for the careful combination with other concepts<br />
on an „at need‟ basis, to conceptualize findings <strong>and</strong> guide further research. I consider<br />
CoP to be an analytical unit fitting these purposes, being adequate for shedding light on<br />
collaborative <strong>work</strong> <strong>and</strong> the use <strong>of</strong> collaboration tools within teams <strong>and</strong> between the team<br />
<strong>and</strong> other project stakeholders, <strong>and</strong> allowing for a combination with the „st<strong>and</strong>ard<br />
repertoire‟ <strong>of</strong> CSCW (e.g. the concepts <strong>of</strong> awareness (Dourish <strong>and</strong> Bellotti 1992;<br />
Gutwin <strong>and</strong> Greenberg 2002) <strong>and</strong> coordination (Carstensen <strong>and</strong> Schmidt 2002 (1999);<br />
Schmidt <strong>and</strong> Simone 1996)).<br />
At the point in the PhD <strong>work</strong> where attention was set on <strong>reflection</strong> there was a need for<br />
theory shedding light on the process <strong>of</strong> <strong>reflection</strong> specifically. Strauss‟ concept <strong>of</strong><br />
project trajectories (Strauss 1993) <strong>and</strong> the model <strong>of</strong> the reflective process (Boud et al.<br />
1985a) could be used as tools for analyzing the reflective process in a team. A strength<br />
<strong>of</strong> the model <strong>of</strong> the reflective process is that it provides support for design by outlining<br />
specific elements that may be included <strong>and</strong> supported in a <strong>reflection</strong> process. <strong>The</strong> choice<br />
<strong>of</strong> drawing on SE project retrospectives approaches from industry in the PhD <strong>work</strong> as a<br />
starting point for implementing the reflective processes in the project teams emerged<br />
from personal communication with a colleague engaged in such practices.<br />
<strong>The</strong> combination <strong>of</strong> the above theories <strong>and</strong> approaches led to a general research focus<br />
on the role <strong>of</strong> representations in <strong>reflection</strong> <strong>and</strong> a need to theoretically frame the process<br />
<strong>of</strong> <strong>reflection</strong> in the context <strong>of</strong> the team‟s overall activity. Activity theory was an option,<br />
having a strength in accounting for the role <strong>of</strong> tools (“Its richness for CSCW lies in<br />
principle in its approach to technology as a mediator <strong>of</strong> human activity, which in a<br />
dialectic relationship with the cultural world produces activity.” (R<strong>and</strong>all et al. 2007,<br />
p.90). However, the role <strong>of</strong> representations, internal <strong>and</strong> external, is more explicitly<br />
addressed in DCog. DCog allows for a combination <strong>of</strong> a social <strong>and</strong> cognitive<br />
perspective on collaborative <strong>work</strong>, being a frame<strong>work</strong> “capable <strong>of</strong> capturing cognitive<br />
activities as embodied <strong>and</strong> situated within the context in which they occur: social <strong>and</strong><br />
organizational” (Rogers <strong>and</strong> Ellis 1994). Combining this with theory accounting for the<br />
process <strong>of</strong> <strong>reflection</strong>, I had a frame<strong>work</strong> adequate for analyzing the case forming the<br />
basis <strong>of</strong> P7 <strong>and</strong> P8. Concepts from DCog provided a language for the <strong>reflection</strong> model<br />
75
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
in P8, enabling the generalization <strong>of</strong> results from the case studies to a model <strong>of</strong><br />
<strong>reflection</strong> in project based <strong>learning</strong>.<br />
76
8 Conclusion <strong>and</strong> further <strong>work</strong><br />
This thesis has approached the challenge <strong>of</strong> supporting <strong>learning</strong> in SE student projects<br />
by focusing on day-to-day project <strong>work</strong> <strong>and</strong> retrospective <strong>reflection</strong> on that <strong>work</strong>. <strong>The</strong><br />
underlying idea is that within project based <strong>learning</strong>, doing real <strong>work</strong> is the driving<br />
force for <strong>learning</strong>. <strong>The</strong> students should accordingly be allowed to keep their focus on<br />
the challenges <strong>of</strong> collaborative <strong>work</strong> <strong>and</strong> gain experience from addressing them as<br />
S<strong>of</strong>tware Engineers. In the thesis, lightweight collaboration tools have been shown to be<br />
valuable in day-to-day SE <strong>work</strong> in different ways <strong>and</strong> for different purposes. Also, the<br />
lightweight collaboration tools have been shown to be potential resources for<br />
retrospective <strong>reflection</strong>, in SE student projects <strong>and</strong> in project based <strong>learning</strong> more<br />
generally, helping to integrate daily <strong>work</strong> <strong>and</strong> <strong>reflection</strong> on that <strong>work</strong> (Figure 17).<br />
Figure 17: Support for <strong>learning</strong> in S<strong>of</strong>tware Engineering student projects as<br />
approached in this thesis<br />
<strong>The</strong> findings <strong>of</strong> the thesis point to a need for a conscious <strong>and</strong> reflective use <strong>of</strong><br />
collaboration tools in SE project courses; among students, course organizers <strong>and</strong><br />
supervisors. <strong>The</strong> thesis illustrates how h<strong>and</strong>ling <strong>of</strong> project challenges by project teams,<br />
frequently involving cross-community collaboration, is closely connected to the use <strong>of</strong><br />
collaboration tools. <strong>The</strong> students would accordingly benefit from insights about the<br />
strengths <strong>and</strong> weaknesses <strong>of</strong> various collaboration tools <strong>and</strong> tool usage, contextualized<br />
77
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
in the typical challenges <strong>of</strong> SE <strong>and</strong> collaborative project <strong>work</strong>. Some initial ideas for<br />
how this could be achieved include: training course staff to provide advice on this topic;<br />
conducting introductory lectures; arranging <strong>work</strong>shops where students from different<br />
groups compare their experiences, <strong>and</strong> addressing the topic <strong>of</strong> collaboration tool use in<br />
retrospective <strong>reflection</strong> <strong>work</strong>shops during the semester. How the topic <strong>of</strong> collaboration<br />
tool use may be more consciously addressed in SE project courses deserves systematic<br />
<strong>and</strong> empirical research by SE education practitioners.<br />
<strong>The</strong> thesis shows some cases <strong>of</strong> collaboration tool use in which cross-community<br />
collaboration entails a relatively high degree <strong>of</strong> participation in the other community,<br />
for example when a project customer gets his own page in a project team‟s wiki, or in<br />
the case <strong>of</strong> OSS participation by a project team, through an internet developer forum. In<br />
the context <strong>of</strong> SE student projects, further research could address the questions <strong>of</strong> when<br />
such participation in the other community is appropriate <strong>and</strong> what it takes for it to<br />
succeed, specifically addressing the role <strong>of</strong> collaboration tools. <strong>The</strong> concepts <strong>of</strong> CoP,<br />
brokering <strong>and</strong> boundary objects might be useful in the analysis. <strong>The</strong> research could be<br />
done within SE education, TEL or CSCW.<br />
Another task for further research in SE education is to adapt the retrospective <strong>reflection</strong><br />
<strong>work</strong>shop approach outlined in the thesis, with or without the use <strong>of</strong> historical data in<br />
collaboration tools, to the needs <strong>of</strong> specific courses. This should be considered <strong>and</strong><br />
conducted as design research <strong>and</strong> not merely „use‟ <strong>of</strong> research results. <strong>The</strong> outcomes<br />
should be improvement <strong>of</strong> the specific project courses as well as guidelines <strong>and</strong><br />
templates for the organizing <strong>of</strong> retrospective <strong>reflection</strong> <strong>work</strong>shops in SE student<br />
projects. <strong>The</strong> more systematic use <strong>of</strong> issue trackers in retrospective <strong>reflection</strong> in a larger<br />
number <strong>of</strong> teams should be explored. An important aspect <strong>of</strong> adapting the approaches to<br />
specific courses is the integration <strong>of</strong> <strong>reflection</strong> into the development processes (e.g. as<br />
done in SCRUM <strong>and</strong> other agile development methodologies). This implies making<br />
<strong>reflection</strong> part <strong>of</strong> a process improvement effort <strong>and</strong> conducting <strong>reflection</strong> <strong>work</strong>shops<br />
several times during the project. In each <strong>work</strong>shop, not only should the teams<br />
collaboratively determine what are main lessons learned, they should define a set <strong>of</strong><br />
action items to be implemented in their subsequent <strong>work</strong>. Accordingly, empirical<br />
research on <strong>reflection</strong> in the projects should investigate the teams‟ actual use <strong>of</strong>, <strong>and</strong><br />
benefit from, the outcomes <strong>of</strong> <strong>reflection</strong>. Comparing data from several <strong>reflection</strong><br />
<strong>work</strong>shops in a project might provide useful data in this respect. This could be<br />
supplemented by longitudinal observation <strong>of</strong> <strong>work</strong> in the team with a focus on the<br />
follow-up <strong>of</strong> action items.<br />
In PBL in other domains <strong>of</strong> education, a similar way <strong>of</strong> organizing retrospective<br />
<strong>reflection</strong> can be adapted <strong>and</strong> tried out. <strong>The</strong> results <strong>of</strong> such studies should be best<br />
78
Conclusion <strong>and</strong> recommendations for further <strong>work</strong><br />
practices <strong>and</strong> guidelines for PBL in the particular domains, <strong>and</strong> should also contribute to<br />
the development <strong>of</strong> these <strong>work</strong> domains beyond educational settings.<br />
<strong>The</strong> insights from this thesis about the potential <strong>of</strong> historical data in collaboration tools<br />
to aid retrospective <strong>reflection</strong> should be developed into a contribution to the s<strong>of</strong>tware<br />
engineering community, as the rationale for the proposed use <strong>of</strong> historical data<br />
originates in SE research <strong>and</strong> the needs <strong>of</strong> SE pr<strong>of</strong>essionals. <strong>The</strong> empirical data on the<br />
use <strong>of</strong> Trac in retrospective <strong>reflection</strong>, as well as new studies in which a similar<br />
approach is introduced in a larger number <strong>of</strong> student projects, should result in<br />
publications directed at the SE community with a focus on how the proposed approach<br />
meets the needs <strong>of</strong> S<strong>of</strong>tware Engineering. Exposing the research to the s<strong>of</strong>tware<br />
engineering community (<strong>and</strong> not just SE education) would help ensure that the use <strong>of</strong><br />
historical data to aid retrospective <strong>reflection</strong> in SE student projects become not only an<br />
improvement <strong>of</strong> education practices but, in line with the PBL core idea <strong>of</strong> doing „real<br />
<strong>work</strong>‟, an improvement <strong>of</strong> SE practice.<br />
<strong>The</strong> thesis outlined the first steps <strong>of</strong> an investigation <strong>of</strong> the connection between day-today<br />
use <strong>of</strong> lightweight collaboration tools, tool features <strong>and</strong> the potential <strong>of</strong> the tools to<br />
aid retrospective <strong>reflection</strong>. Further CSCW research along this line should aim for the<br />
development <strong>of</strong> a frame<strong>work</strong> that can be used by project teams <strong>and</strong> organizers as well as<br />
designers <strong>of</strong> collaboration tools to consider the role <strong>of</strong> collaboration tools in<br />
retrospective <strong>reflection</strong>. <strong>The</strong> development <strong>of</strong> the frame<strong>work</strong> would require a survey <strong>of</strong><br />
state-<strong>of</strong>-the-art research on day-to-day tool use in teams combined with empirical<br />
studies addressing the connection between day-to-day tool use <strong>and</strong> the use <strong>of</strong> the same<br />
tools in retrospective <strong>reflection</strong>. Again, SE student projects may be used as test beds.<br />
<strong>The</strong> <strong>reflection</strong> model proposed in the thesis can be used to inform further research on<br />
reflective <strong>learning</strong> <strong>and</strong> the role <strong>of</strong> collaboration tools in supporting day-to-day<br />
collaborative <strong>work</strong> <strong>and</strong> <strong>reflection</strong> on that <strong>work</strong>. Based on the results <strong>of</strong> such research,<br />
the model can be further developed <strong>and</strong> refined. I propose three research directions that<br />
can be pursued independently or in combination:<br />
Firstly, the model can be used as a basis for exploring <strong>and</strong> comparing various types <strong>of</strong><br />
use <strong>of</strong> historical data in day-to-day <strong>work</strong>, including retrospective <strong>reflection</strong>. This<br />
research should identify how current approaches <strong>and</strong> solutions to the gathering <strong>and</strong> use<br />
<strong>of</strong> data for one purpose (e.g. coordination) – or the gathered data themselves - could be<br />
applied for another purpose (e.g. <strong>reflection</strong>) within the <strong>work</strong> practice. <strong>The</strong> research<br />
should draw on state-<strong>of</strong>-the-art research within SE <strong>and</strong> CSCW <strong>and</strong> result in<br />
contributions (to the same communities) with practical applicability for the organization<br />
<strong>of</strong>, <strong>and</strong> tool support for, collaborative project <strong>work</strong>. Also, the research should lead to a<br />
79
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
refinement <strong>of</strong> the <strong>reflection</strong> model with respect to the role <strong>of</strong> collaboration tools in<br />
integrating different aspects <strong>of</strong> the <strong>work</strong> practice.<br />
Secondly, the model can be used to frame studies <strong>of</strong> retrospective <strong>reflection</strong> across<br />
different domains. <strong>The</strong> studies may be based on use <strong>of</strong> the timeline <strong>and</strong> satisfaction<br />
curve technique in a way similar to the <strong>work</strong>shops <strong>of</strong> the SE student projects described<br />
in the thesis, but they may also be based other approaches instantiating the model in<br />
different ways (e.g. with individual <strong>and</strong>/or collaborative <strong>reflection</strong> efforts, use <strong>of</strong> various<br />
representations to aid <strong>reflection</strong>, <strong>and</strong> with the possible use <strong>of</strong> historical data from one or<br />
more collaboration tools used in daily <strong>work</strong>). <strong>The</strong> different instantiations <strong>of</strong> the <strong>work</strong><strong>reflection</strong>-<strong>learning</strong><br />
<strong>cycle</strong> should be used to shed light on similarities <strong>and</strong> differences<br />
between domains <strong>and</strong> between approaches, <strong>and</strong> thus serve as a basis for refinement <strong>of</strong><br />
the model. Such a refinement would increase the value <strong>of</strong> the model as an aid to<br />
designing for the <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in the specific case, helping outline<br />
options <strong>and</strong> unveiling the potential for supporting <strong>learning</strong> by use <strong>of</strong> <strong>reflection</strong> activities<br />
<strong>and</strong> data originating in the <strong>work</strong> process.<br />
Thirdly, <strong>and</strong> related to the previous point, to extend the value <strong>of</strong> the <strong>reflection</strong> model, it<br />
should be elaborated to focus more strongly on the use <strong>of</strong> the results <strong>of</strong> <strong>reflection</strong> as an<br />
aid to the <strong>work</strong> practice in question - within the project, team <strong>and</strong>/or organization. In<br />
other words, the model should be refined as a tool for outlining core elements <strong>of</strong> process<br />
improvement as well as the connection between project <strong>learning</strong> <strong>and</strong> organizational<br />
<strong>learning</strong>. <strong>The</strong> research needed for this refinement should partially be empirically based,<br />
exploring specific cases <strong>and</strong> trying out solutions (both <strong>of</strong> which can be informed by the<br />
<strong>reflection</strong> model). <strong>The</strong> research suggested above, to adapt the use <strong>of</strong> retrospective<br />
<strong>reflection</strong> <strong>work</strong>shops to specific SE project courses, can be considered part <strong>of</strong> this<br />
effort. Also, the research should draw on the literature <strong>of</strong> SE practice <strong>and</strong> research,<br />
organizational <strong>learning</strong>, knowledge management <strong>and</strong> project management, to utilize<br />
state-<strong>of</strong>-the-art knowledge <strong>of</strong> process improvement <strong>and</strong> organizational <strong>learning</strong>.<br />
In conclusion, this thesis brings forward the underst<strong>and</strong>ing <strong>of</strong> <strong>work</strong> <strong>and</strong> <strong>learning</strong> in SE<br />
student teams <strong>and</strong> how <strong>work</strong> <strong>and</strong> <strong>learning</strong> can be bridged by the use <strong>of</strong> collaboration<br />
tools aiding retrospective <strong>reflection</strong>. Further research is needed to explore the potential<br />
<strong>of</strong> collaboration tools to support such <strong>reflection</strong> in project teams. This research should<br />
consider the particular characteristics <strong>of</strong> the <strong>work</strong>/<strong>learning</strong> setting (e.g. issues <strong>of</strong> crosscommunity<br />
collaboration), the collaboration tool features <strong>and</strong> usage (e.g. the concerted<br />
use <strong>of</strong> tools), <strong>and</strong> the organization <strong>of</strong> the reflective activity (e.g. the use <strong>of</strong> techniques to<br />
visualize the project process).<br />
80
9 References<br />
Abran, A., Moore, J. W., Bourque, P., <strong>and</strong> Dupnis, R. (2004). "Guide to <strong>The</strong> S<strong>of</strong>tware<br />
Engineering Body <strong>of</strong> Knowledge ". City: IEEE.<br />
Ackerman, M. S., <strong>and</strong> Halverson, C. (2004). "Organizational Memory as Objects,<br />
Process, <strong>and</strong> Trajectories: An Examination <strong>of</strong> Organizational Memory in Use."<br />
<strong>Computer</strong> Supported Cooperative Work, 13(2), 155-189.<br />
Akkerman, S., van den Bossche, P., Admiraal, W., Gijselaers, W., Segers, M., Simons,<br />
R.-J., <strong>and</strong> Kirschner, P. A. (2007). "Reconsidering group cognition: From<br />
conceptual confusion to a boundary area between cognitive <strong>and</strong> socio-cultural<br />
perspectives?" Educational Research Review, 2(1), 39-63.<br />
Ar<strong>and</strong>a, J., <strong>and</strong> Venolia, G. "<strong>The</strong> Secret Life Of Bugs: Going Past the Errors <strong>and</strong><br />
Omssions in S<strong>of</strong>tware Repositories " Presented at International Conference on<br />
S<strong>of</strong>tware Engineering (ICSE'09), Vancouver, Canada.<br />
Armarego, J. "Beyond PBL: Preparing Graduates for Pr<strong>of</strong>essional Practice." Presented<br />
at S<strong>of</strong>tware Engineering Education & Training, CSEE&T, Dublin.<br />
B<strong>and</strong>ura, A. (1996). "Social cognitive theory <strong>of</strong> human development", in T. Husen <strong>and</strong><br />
T. N. Postlethwaite, (eds.), International encyclopedia <strong>of</strong> education. Oxford:<br />
Pergamon Press.<br />
Barab, S. A., <strong>and</strong> Squire, K. (2004). "Design-Based Research: Putting a Stake in the<br />
Ground." Journal <strong>of</strong> the Learning Sciences, 13(1), 1-14.<br />
Baron, N. S. (2004). "See you Online: Gender Issues in College Student Use <strong>of</strong> Instant<br />
Messaging." Journal <strong>of</strong> Language <strong>and</strong> Social Psychology, 23(4).<br />
Basili, V. R., <strong>and</strong> Caldiera, G. (1995). "Improving S<strong>of</strong>tware Quality by Reusing<br />
Knowledge <strong>and</strong> Experience." Sloan Management Review, Fall 1995.<br />
Beck, K. (2000). "Embracing change with extreme programming." IEEE <strong>Computer</strong>,<br />
32(10), 70-78.<br />
Berger, P., <strong>and</strong> Luckmann, T. (1966). <strong>The</strong> Social Construction <strong>of</strong> Reality, London:<br />
Penguin Books.<br />
Berglund, A., <strong>and</strong> Eckerdal, A. (2006). "What do CS students try to learn? insights from<br />
a distributed, project-based course in computer systems." <strong>Computer</strong> Science<br />
Education, 16(3), 11.<br />
Bjørnson, F. O., Wang, A. I., <strong>and</strong> Arisholm, E. (2009). "Improving the effectiveness <strong>of</strong><br />
root cause analysis in post mortem analysis: A controlled experiment."<br />
Information <strong>and</strong> S<strong>of</strong>tware Technology, 51, 150-161.<br />
Bl<strong>and</strong>ford, A., <strong>and</strong> Furniss, D. (2006). "DiCoT: A Methodology for Applying<br />
Distributed Cognition to the Design <strong>of</strong> Team<strong>work</strong>ing Systems", in S. W. Gilroy<br />
<strong>and</strong> M. D. Harrison, (eds.), DSVIS 2005. Berlin Heidelberg: Springer-Verlag.<br />
81
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Blumer, H. (1954). "What Is Wrong with Social <strong>The</strong>ory?" American Sociological<br />
Review, 19(1), 3-10.<br />
Blumer, H. (1986). Symbolic Interactionism: Perspective <strong>and</strong> Method, Berkley:<br />
University <strong>of</strong> California Press.<br />
Boehm, B. (1988). "A spiral model <strong>of</strong> s<strong>of</strong>tware development <strong>and</strong> enhancement." IEEE<br />
<strong>Computer</strong>, 21(5), 61-72.<br />
Boud, D., Keogh, R., <strong>and</strong> Walker, D. (1985a). "Promoting Reflection in Learning: a<br />
Model", in D. Boud, R. Keogh, <strong>and</strong> D. Walker, (eds.), Reflection: Turning<br />
Experience into Learning. RoutledgeFalmer, pp. 18-40.<br />
Boud, D., Keogh, R., <strong>and</strong> Walker, D. (1985b). Reflection: Turning Experience into<br />
Learning: RoutledgeFalmer.<br />
Bowen, G. A. (2006). "Grounded <strong>The</strong>ory <strong>and</strong> Sensitizing Concepts." International<br />
Journal <strong>of</strong> Qualitative Methods, 5(3).<br />
Brady, A., Johnson, R. R., <strong>and</strong> Wallace, C. (2006). "<strong>The</strong> intersecting futures <strong>of</strong> technical<br />
communication <strong>and</strong> s<strong>of</strong>tware engineering: Forging a multi-disciplinary alliance."<br />
Technical Communication, 53(3).<br />
Bratteteig, T., <strong>and</strong> Gregory, J. "Human Action in Context. A Discussion <strong>of</strong> <strong>The</strong>ories for<br />
Underst<strong>and</strong>ing Use <strong>of</strong> IT." Presented at IRIS'22, Jyväskylä.<br />
Brown, A. L. (1992). "Design Experiments: <strong>The</strong>oretical <strong>and</strong> Methodological Challenges<br />
in Creating Complex Interventions in Classroom Settings." <strong>The</strong> Journal <strong>of</strong> the<br />
Learning Sciences, 2(2), 141-178.<br />
Brown, J. S., Collins, A., <strong>and</strong> Duguid, P. (1989). "Situated Cognition <strong>and</strong> the Culture <strong>of</strong><br />
Learning." Educational Researcher, 18(1), 32-42.<br />
Bruner, J. (1990). Acts <strong>of</strong> Meaning: Harvard University Press.<br />
Bråten, S. (1981). "Quality <strong>of</strong> Interaction <strong>and</strong> Participation. On Model Power in<br />
Industrial Democracy <strong>and</strong> <strong>Computer</strong> Net<strong>work</strong>s", in G. E. Lasker, (ed.), Applied<br />
Systems <strong>and</strong> Cybernetics. pp. 191-200.<br />
Buhl, H., <strong>and</strong> Richter, A. (2004). "Downplaying Model Power in IT Project Work."<br />
Economic <strong>and</strong> Industrial Democracy, 25.<br />
Burge, J., <strong>and</strong> Brown, D. C. "SEURAT: integrated rationale management." Presented at<br />
ICSE'08, Leipzig, Germany.<br />
Burge, J., <strong>and</strong> Kiper, J. "Capturing Collaborative Design Decisions <strong>and</strong> Rationale."<br />
Presented at Design, Computing, <strong>and</strong> Cognition, Atlanta, Georgia, USA.<br />
Burnett, R. E. (2001). Technical Communication, Fifth Edtion, Boston, MA, USA:<br />
Thomson Heinle.<br />
Bygstad, B., Krogstie, B. R., <strong>and</strong> Grønli, T. M. (2009). "Learning from achievement:<br />
scaffolding student projects in s<strong>of</strong>tware engineering " International Journal <strong>of</strong><br />
Net<strong>work</strong>ed <strong>and</strong> Virtual Organizations, 6(2).<br />
Carstensen, P. H., <strong>and</strong> Schmidt, K. (2002 (1999)). "<strong>Computer</strong> Supported Cooperative<br />
Work: New Challenges to Systems Design", in K. Itoh, (ed.), H<strong>and</strong>book <strong>of</strong><br />
Human Factors/Ergonomics. Tokyo: Asakura Publishing.<br />
Cherry, S., <strong>and</strong> Robillard, P. N. (2008). "<strong>The</strong> social side <strong>of</strong> s<strong>of</strong>tware engineering - A<br />
real ad hoc collaboration net<strong>work</strong>." International Journal <strong>of</strong> Human-<strong>Computer</strong><br />
Studies, 66, 495-505.<br />
Coates, H., James, R., <strong>and</strong> Baldwin, G. (2005). "A critical examination <strong>of</strong> the effects <strong>of</strong><br />
<strong>learning</strong> management systems on university teaching <strong>and</strong> <strong>learning</strong>." Tertiary<br />
Education <strong>and</strong> Management 11, 17.<br />
82
References<br />
Cobb, P. (1994). "Where is the Mind? Constructivist <strong>and</strong> Sociocultural Perspectives on<br />
Mathematical Development." Educational Researcher, 23(7), 8.<br />
Cobb, P., Confrey, J., diSessa, A., Lehrer, R., <strong>and</strong> Schauble, L. (2003). "Design<br />
Experiments in Educational Research." Educational Researcher, 32(1), 9-13.<br />
Cockburn, A. (2000). "Selecting a Project's Methodology." IEEE S<strong>of</strong>tware<br />
(July/August), 64-71.<br />
Cockburn, A. (2006). Agile S<strong>of</strong>tware Development: <strong>The</strong> Cooperative Game, 2nd<br />
edition: Addison-Wesley Pr<strong>of</strong>essional.<br />
Collier, B., DeMarco, T., <strong>and</strong> Fearey, P. (1996). "A defined process for project post<br />
mortem review." IEEE S<strong>of</strong>tware, 13(4), 8.<br />
Collins, A. (1992). "Toward a Design Science <strong>of</strong> Education", in E. Scanlon <strong>and</strong> T.<br />
O'Shea, (eds.), New Directions in Educational Technology. Springer-Verlag.<br />
Conklin, E. J., <strong>and</strong> Yakemovic, K. B. (1991). "A Process-Oriented Approach to Design<br />
Rationale." Human-<strong>Computer</strong> Interaction, 6, 357-391.<br />
Cress, U., <strong>and</strong> Kimmerle, J. (2008). "A systemic <strong>and</strong> cognitive view on collaborative<br />
knowledge building in wikis." <strong>Computer</strong>-Supported Collaborative Learning, 3,<br />
105-122.<br />
Cubranic, D., Murphy, G. C., Singer, J. S., <strong>and</strong> Booth, K. S. "Learning from project<br />
history: a case study for s<strong>of</strong>tware development." Presented at CSCW '04,<br />
Chicago, USA.<br />
Daradoumis, T., <strong>and</strong> Marques, M. (2002). "Distributed Cognition in the Context <strong>of</strong><br />
Virtual Collaborative Learning." Journal <strong>of</strong> Interactive Learning Research,<br />
13(1/2), 135-148.<br />
Decker, B., Ras, E., Rech, J., Jaubert, P., <strong>and</strong> Rieth, M. (2007). "Wiki-Based<br />
Stakeholder Participation in Requirements Engineering." IEEE S<strong>of</strong>tware,<br />
2007(March/April).<br />
Dekel, U., <strong>and</strong> Herbsleb, J. D. "Pushing Relevant Artifact Annotationsin Collaborative<br />
S<strong>of</strong>tware Development." Presented at CSCW'08, San Diego, USA.<br />
Derby, E., Larsen, D., <strong>and</strong> Schwaber, K. (2006). Agile Retrospectives. Making Good<br />
Teams Great: Pragmatic Bookshelf.<br />
Dewey, J. (1933). How we think. A restatement <strong>of</strong> the relation <strong>of</strong> reflective thinking to<br />
the educative process Boston: D. C. Heath.<br />
Dewey, J. (1978 (1910)). "A short catechism concerning truth", in J. A. Boydston, (ed.),<br />
John Dewey. <strong>The</strong> middle <strong>work</strong>s 1899-1924.: Southern Illinois University Press.<br />
Dewey, J. (2008 (1909)). Studies in Logical <strong>The</strong>ory: Bibliobazaar.<br />
Dingsøyr, T. (2005). "Postmortem reviews: purpose <strong>and</strong> approaches in s<strong>of</strong>tware<br />
engineering." Information <strong>and</strong> S<strong>of</strong>tware Technology, 47, 293-303.<br />
Dohn, N. B. (2009). "Web 2.0: Inherent tensions <strong>and</strong> evident challenges for education."<br />
<strong>Computer</strong>-Supported Collaborative Learning, 4, 343-363.<br />
Dourish, P. "Implications for Design." Presented at CHI, Montreal, Quebec, Canada.<br />
Dourish, P., <strong>and</strong> Bellotti, V. "Awareness <strong>and</strong> coordination in shared <strong>work</strong>spaces."<br />
Presented at ACM CSCW, Toronto, Ontario, Canada<br />
Duim, L. v. d., Jesper, A., <strong>and</strong> Sinnema, M. "Good practices for Educational S<strong>of</strong>tware<br />
Engineering Projects." Presented at 29th International Conference on S<strong>of</strong>tware<br />
Engineering (ICSE'07), Minneapolis, USA.<br />
Durkee, D., Brant, S., Nevin, P., Odell, A., Williams, G., Melomey, D., Roberts, H.,<br />
Imafidion, C., Perryman, R., <strong>and</strong> Lopes, a. (2009). "Implementing E-Learning<br />
83
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
<strong>and</strong> Web 2.0 Innovation: Didactical Scenarios <strong>and</strong> Practical Implications."<br />
Industry <strong>and</strong> Higher Education, 23(4), 8.<br />
Dybå, T., <strong>and</strong> Dingsøyr, T. (2008). "Empirical studies <strong>of</strong> agile s<strong>of</strong>tware development: A<br />
systematic review." Information <strong>and</strong> S<strong>of</strong>tware Technology, 2008(50), 833-859.<br />
Engeström, Y. (1987). Learning by exp<strong>and</strong>ing. An activity-theoretical approach to<br />
developmental research: Orienta-Konsultit Oy, Helsinki.<br />
Engeström, Y. (2000). "From individual action to collective activity <strong>and</strong> back:<br />
developmental <strong>work</strong> research as an interventionist methodology", in P. Luff, J.<br />
Hindmarsh, <strong>and</strong> C. Heath, (eds.), Workplace Studies - Recovering Work Practice<br />
<strong>and</strong> Informing System Design. Cambridge University Press.<br />
Eraut, M. (2000). "Non-formal <strong>learning</strong> <strong>and</strong> tacit knowledge in pr<strong>of</strong>essional <strong>work</strong>."<br />
British Journal <strong>of</strong> Educational Psychology, 70, 24.<br />
Eraut, M., <strong>and</strong> Hirsch, W. (2007). <strong>The</strong> Significance <strong>of</strong> Workplace Learning for<br />
Individuals, Groups <strong>and</strong> Organisations. Oxford & Cardiff Universities.<br />
Fitzpatrick, G. (2003). <strong>The</strong> Locales Frame<strong>work</strong>: Underst<strong>and</strong>ing <strong>and</strong> Designing for<br />
Wicked Problems: Kluwer Academic Publishers.<br />
Fitzpatrick, G., Marshall, P., <strong>and</strong> Phillips, A. "CVS integration with notification <strong>and</strong><br />
chat: lightweight s<strong>of</strong>tware team collaboration." Presented at CSCW'06, Banff,<br />
Alberta, Canada.<br />
Fuchs-Kittowski, F., <strong>and</strong> Köhler, A. "Wiki Communities in the Context <strong>of</strong> Work<br />
Processes." Presented at 2005 international symposium on Wikis WikiSym '05<br />
Gal, U., Youngjin, Y., <strong>and</strong> Bol<strong>and</strong>, R. (2005). "<strong>The</strong> dynamics <strong>of</strong> boundary objects,<br />
social infrastructures <strong>and</strong> social identities."<br />
Germain, E., <strong>and</strong> Robillard, P. N. (2004). "Engineering-based processes <strong>and</strong> agile<br />
methodologies for s<strong>of</strong>tware development: a comparative case study." <strong>The</strong><br />
Journal <strong>of</strong> Systems <strong>and</strong> S<strong>of</strong>tware, 75, 17-27.<br />
Gleaves, a., Walker, C., <strong>and</strong> Grey, J. (2007). "Using digital <strong>and</strong> paper diaries for<br />
<strong>learning</strong> <strong>and</strong> assessment purposes in higher education: a comparative study <strong>of</strong><br />
feasibility <strong>and</strong> reliability." Assessment & Evaluation in Higher Education, 32(6),<br />
631-643.<br />
Grinter, R. E., <strong>and</strong> Palen, L. "Instant Messaging in Teen Life." Presented at CSCW'02,<br />
New Orelans, Louisiana, USA.<br />
Gutwin, C., <strong>and</strong> Greenberg, S. (2002). "A Descriptive Frame<strong>work</strong> <strong>of</strong> Workspace<br />
Awareness for Real-Time Groupware." <strong>Computer</strong> Supported Cooperative Work<br />
11(3-4), 411-446.<br />
Gutwin, C., Penner, R., <strong>and</strong> Schneider, K. "Group Awareness in Distributed S<strong>of</strong>tware<br />
Development." Presented at CSCW'04, Chicago, Illinois, USA.<br />
Gutwin, C., Penner, R., <strong>and</strong> Schneider, K. "Knowledge sharing in s<strong>of</strong>tware engineering:<br />
Group awareness in distributed s<strong>of</strong>tware development " Presented at CSCW'04,<br />
Chicago, Illinois, USA.<br />
Halverson, C. (2002). "Activity <strong>The</strong>ory <strong>and</strong> Distributed Cognition: Or What Does<br />
CSCW Need to DO with <strong>The</strong>ories?" <strong>Computer</strong> Supported Cooperative Work,<br />
11, 243-267.<br />
Helle, L., Tynjälä, P., <strong>and</strong> Olkinuora, E. (2006). "Project-based <strong>learning</strong> in postsecondary<br />
education - theory, practice <strong>and</strong> rubber sling shots." Higher<br />
Education, 51, 287-314.<br />
84
References<br />
Herbsleb, J. D., Mockus, A., Finholt, T. A., <strong>and</strong> Grinter, R., E. "An Empirical Study <strong>of</strong><br />
Global S<strong>of</strong>tware Development: Distance <strong>and</strong> Speed." Presented at International<br />
Conference on S<strong>of</strong>tware Engineering, Toronto, Ontario, Canada.<br />
Hickerson, C. A., <strong>and</strong> Giglio, M. (2009). "Instant Messaging Between Students <strong>and</strong><br />
Faculty: A Tool for Increasing Student-Faculty Interaction." International<br />
Journal on E-<strong>learning</strong>, 8(1), 71-88.<br />
Hildrum, J. M. (2009). "Sharing Tacit Knowledge Online: A Case Study <strong>of</strong> e-Learning<br />
in Cisco's Net<strong>work</strong> <strong>of</strong> System Integrator Partner Firms." Industry <strong>and</strong><br />
Innovation, 16(2).<br />
Huang, A. H., <strong>and</strong> Yen, D. C. (2003). "Usefulness <strong>of</strong> instant messaging among young<br />
users: Social vs. <strong>work</strong> perspective." Human Systems Management, 22(2), 63-72.<br />
Hutchins, E. (1995). Cognition in the Wild, Cambridge, Massachusetts: <strong>The</strong> MIT Press.<br />
Idris, Y., <strong>and</strong> Wang, Q. (2009). "Affordances <strong>of</strong> Facebook for <strong>learning</strong>." International<br />
Journal <strong>of</strong> Continuing Engineering Education <strong>and</strong> Life-Long Learning<br />
(IJCEELL), 19(2/3).<br />
IEEE. (1990). "IEEE St<strong>and</strong>ard Glossary <strong>of</strong> S<strong>of</strong>tware Engineering Terminology"std<br />
610.12-1990. City: IEEE.<br />
Isaacs, E., Camm, C., Schiano, D. J., Walendowski, A., <strong>and</strong> Whittaker, S.<br />
"Characterizing Instant Messaging from Recorded Logs." Presented at CHI<br />
2002.<br />
Isaacs, E., Walendowski, A., Whittaker, S., Schiano, D. J., <strong>and</strong> Kamm, C. "<strong>The</strong><br />
Character, Functions, <strong>and</strong> Styles <strong>of</strong> Instant Messaging in the Workplace."<br />
Presented at CSCW'02, New Orleans, Louisiana, USA.<br />
Jonassen, D. H. (1995). "<strong>Computer</strong>s as Cognitive Tools: Learning with Technology,<br />
Not from Technology." Journal <strong>of</strong> Computing in Higher Education, 6(2), 40-73.<br />
Kaptelinin, V., <strong>and</strong> Nardi, B. A. (2006). Acting with Technology. Activity <strong>The</strong>ory <strong>and</strong><br />
Interaction Design., Cambridge, Massachusetts: MIT Press.<br />
Kasi, V., Keil, M., Mathiassen, L., <strong>and</strong> Pedersen, K. (2008). "<strong>The</strong> post mortem paradox:<br />
a Delphi study <strong>of</strong> IT specialist perceptions." European Journal <strong>of</strong> Information<br />
Systems, 17, 62-78.<br />
Keil, M., Mann, J., <strong>and</strong> Rai, A. (2000). "Why S<strong>of</strong>tware Projects Escalate: An Empirical<br />
Analysis <strong>and</strong> Test <strong>of</strong> Four <strong>The</strong>oretical Models." MIS Quarterly, 24(4).<br />
Kelly, A. E. (2003). "Research as design: <strong>The</strong> role <strong>of</strong> design in educational research."<br />
Educational Researher, 32(1), 3-4.<br />
Kelly, A. E., Baek, J. Y., Lesh, R. A., <strong>and</strong> Bannan-Ritl<strong>and</strong>, B. (2008a). "Enabling<br />
Innovations in Education <strong>and</strong> Systematizing their Impact", in A. E. Kelly, R. A.<br />
Lesh, <strong>and</strong> J. Y. Baek, (eds.), H<strong>and</strong>book <strong>of</strong> Design Research Methods in<br />
Education. New York: Routledge.<br />
Kelly, A. E., Lesh, R. A., <strong>and</strong> Baek, J. Y. (2008b). H<strong>and</strong>book <strong>of</strong> Design Research<br />
Methods in Education: Routledge.<br />
Kerth, N. (2001). Project Retrospectives: A H<strong>and</strong>book for Team Reviews Dorset House<br />
Publishing Company.<br />
Kim, B., <strong>and</strong> Reeves, T. C. (2007). "Reframing research on <strong>learning</strong> with technology: in<br />
search <strong>of</strong> the meaning <strong>of</strong> cognitive tools." Instructional Science, 35, 207-256.<br />
Kim, D., <strong>and</strong> Lee, S. (2002). "Designing Collaborative Reflection Support Tools in e-<br />
project Based Learning Environment." Journal <strong>of</strong> Interactive Learning<br />
Research, 13(4), 375-392.<br />
85
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Kirschner, P. A., <strong>and</strong> Erkens, G. (2006). "Cognitive tools <strong>and</strong> mindtools for<br />
collaborative <strong>learning</strong>." Journal <strong>of</strong> Educational Computing Research, 35(2),<br />
199-209.<br />
Klein, H. K., <strong>and</strong> Myers, M. M. (1999). "A Set <strong>of</strong> Principles for Conducting <strong>and</strong><br />
Evaluating Interpretive Field Studies in Information Systems." MIS Quarterly,<br />
23(1), 67-94.<br />
Kolb, D. A., <strong>and</strong> Fry, R. (1975). "Towards an applied theory <strong>of</strong> experiental <strong>learning</strong>", in<br />
C. L. Cooper, (ed.), <strong>The</strong>ories <strong>of</strong> Group Processes. London: John Wiley, pp. 33-<br />
58.<br />
Kolb, D. A., Rubin, I. M., <strong>and</strong> McIntyre, J. (1971). "Organizational psychology: an<br />
experiential approach". City: Prentice Hall: Englewood Cliffs, NJ.<br />
Krogstie, B. (2000). Applying Activity <strong>The</strong>ory in Knowledge Management, University <strong>of</strong><br />
Oslo, Oslo.<br />
Krogstie, B. "Power through brokering. OSS participation in SE projects." Presented at<br />
International Conference on S<strong>of</strong>tware Engineering (ICSE) 2008, Leipzig.<br />
Krogstie, B., <strong>and</strong> Bygstad, B. "Cross-Community Collaboration <strong>and</strong> Learning in<br />
Customer-Driven S<strong>of</strong>tware Engineering Student Projects." Presented at<br />
Twentieth Conference on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training<br />
(CSEE&T), Dublin.<br />
Krogstie, B., <strong>and</strong> Divitini, M. "Practice-Based Learning as Mobile Learning: <strong>The</strong> Role<br />
<strong>of</strong> Boundary Objects." Presented at Mobile Learning, Lisbon, Portugal.<br />
Krogstie, B. R. "<strong>The</strong> wiki as an integrative tool in project <strong>work</strong>." Presented at COOP,<br />
Carry-le-Rouet, Provence, France.<br />
Krogstie, B. R. "Do's <strong>and</strong> dont's <strong>of</strong> instant messaging in students' project <strong>work</strong>."<br />
Presented at NOKOBIT 2009, Trondheim, Norway.<br />
Krogstie, B. R. "A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> utilizing<br />
historical data in collaborative tools." Presented at EC-TEL 2009, Nice, France.<br />
Krogstie, B. R. "Using Project Wiki History to Reflect on the Project Process "<br />
Presented at 42nd Hawaii International Conference on System Sciences, Big<br />
Isl<strong>and</strong>, Hawaii.<br />
Krogstie, B. R., <strong>and</strong> Divitini, M. "Shared timeline <strong>and</strong> individual experience:<br />
Supporting retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering teams."<br />
Presented at CSEE&T 2009, Hyderabad.<br />
Krogstie, B. R., <strong>and</strong> Divitini, M. "Supporting Reflection in S<strong>of</strong>tware Development with<br />
Everyday Working Tools." Presented at COOP, Aix-en-Provence, France.<br />
Kruchten, P. (2003). <strong>The</strong> Rational Unified Process: An introduction: Addison-Wesley.<br />
Kuutti, K. (1995). "Activity <strong>The</strong>ory as a Potential Frame<strong>work</strong> for Human-<strong>Computer</strong><br />
Interaction Research", in B. A. Nardi, (ed.), Context <strong>and</strong> Consciousness London:<br />
MIT Press.<br />
Kuutti, K., <strong>and</strong> Kaptelinin, V. "Rethinking cognitive tools: from augmentation to<br />
mediation." Presented at Second International Conference on Cognitive<br />
Technology Humanizing the INformation Age.<br />
Latour, B. (2005). Reassembling the Social. An Introduction to Actor-Net<strong>work</strong>-<strong>The</strong>ory:<br />
Oxford University Press.<br />
Lave, J., <strong>and</strong> Wenger, E. (1991). Situated Learning. Legitimate peripheral<br />
participation, Cambridge: University <strong>of</strong> Cambridge Press.<br />
86
References<br />
Leont'ev, A. N. (1981). Problems <strong>of</strong> the development <strong>of</strong> the mind: Progress Publishers,<br />
Moscow.<br />
Lin, X., Hmelo, C., Kinzer, C. K., <strong>and</strong> Secules, T. J. (1999). "Designing Technology to<br />
Support Reflection." Educational Technology, Research <strong>and</strong> Development,<br />
47(3), 43-62.<br />
Louridas, P. (2006). "Using Wikis in S<strong>of</strong>tware Development." IEEE S<strong>of</strong>tware, 23(2 ).<br />
Lovejoy, T., <strong>and</strong> Grudin, J. "Messaging And Formality: Will IM Follow in the<br />
Footsteps <strong>of</strong> Email?" Presented at INTERACT Zurich, Switzerl<strong>and</strong>.<br />
Lund, A., <strong>and</strong> Smørdal, O. "Is <strong>The</strong>re a Space for the Teacher in a WIKI?" Presented at<br />
Proceedings <strong>of</strong> the 2006 international symposium on Wikis WikiSym '06<br />
Lyytinen, K., <strong>and</strong> Robey, D. (1999). "Learning failure in information systems<br />
development." Information Systems Journal, 9, 17.<br />
Majchrzak, A., Wagner, C., <strong>and</strong> Yates, D. "Corporate Wiki Users: Results <strong>of</strong> a Survey."<br />
Presented at WikiSym, Odense, Denmark.<br />
McCarthy, J., <strong>and</strong> Wright, P. (2004). Technology as Experience: MIT Press.<br />
McMillan, W. W. "What Leading Practitioners Say Should Be Emphasized in Students'<br />
S<strong>of</strong>tware Engineering Projects." Presented at Conference on S<strong>of</strong>tware<br />
Engineering Education <strong>and</strong> Training, New Orleans, Louisiana, USA.<br />
Mead, G. H. (1934). Mind, Self <strong>and</strong> Society, Chicago: <strong>The</strong> University <strong>of</strong> Chicago Press.<br />
Muller, M., Raven, M. E., Kogan, S., Millen, D. R., <strong>and</strong> Carey, K. "Introducing Chat<br />
into Business Organisations: Toward an Instant Messaging Maturity Model."<br />
Presented at GROUP'03, Sanibel Isl<strong>and</strong>, Florida, USA.<br />
Nardi, B. A. "Studying context: a comparison <strong>of</strong> activity theory, situated action models,<br />
<strong>and</strong> distributed cognition." Presented at St.Petersburg International Workshop<br />
on Human-<strong>Computer</strong> Interaction, St.Petersburg, USSR.<br />
Nardi, B. A. (2002). "Coda <strong>and</strong> Response to Christine Halverson." <strong>Computer</strong> Supported<br />
Cooperative Work, 11, 269-275.<br />
Nardi, B. A., Whittaker, S., <strong>and</strong> Bradner, E. "Interaction <strong>and</strong> Outeraction: Instant<br />
Messaging in Action." Presented at CSCW'00, Philadelphia, PA, USA.<br />
Niinimaki, T. "Experiences <strong>of</strong> instant messaging in global s<strong>of</strong>tware development<br />
projects: a multiple case study." Presented at International Conference on<br />
Global S<strong>of</strong>tware Engineering (ICGSE), Piscataway, NJ, USA.<br />
Omoronyia, I., Ferguson, J., Roper, M., <strong>and</strong> Wood, M. (2009). "Using Developer<br />
Activity Data to Enhance Awareness during Collaborative S<strong>of</strong>tware<br />
Development." <strong>Computer</strong> Supported Cooperative Work, 18, 50.<br />
OpenSourceInitiative. (2010). "<strong>The</strong> Open Source Definition". City.<br />
Palincsar, A. S. (1998). "Social Constructivist Perspectives on Teaching <strong>and</strong> Learning."<br />
Annual Review <strong>of</strong> Psychology, 49, 345-75.<br />
Pea, R. (1985). "Beyond Amplification: Using the <strong>Computer</strong> to Reorganize Mental<br />
Functioning." Educational Psychologist, 20(4).<br />
Pea, R. (1994). "Seeing What We Build Together: Distributed Multimedia Learning<br />
Environments for Transformative Communications." <strong>The</strong> Journal <strong>of</strong> the<br />
Learning Sciences, 3(3), 15.<br />
Phillips, D. C. (1995). "<strong>The</strong> Good, the Bad, <strong>and</strong> the Ugly: <strong>The</strong> Many Faces <strong>of</strong><br />
Constructivism." Educational Researcher, 24(7), 5-12.<br />
87
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Pipek, V., <strong>and</strong> Wulf, V. (2007). "From Groupware towards Collaborative<br />
Infrastructures. Panel presentation."European conference on <strong>Computer</strong>-<br />
Supported Cooperative Work. City: Limerick, Irel<strong>and</strong>.<br />
Polanyi, M. (1966). <strong>The</strong> Tacit Dimension: Doubleday & Co.<br />
Poole, W. G. "<strong>The</strong> S<strong>of</strong>ter Side <strong>of</strong> Custom S<strong>of</strong>tware Development: Working with the<br />
Other Players." Presented at Conference on S<strong>of</strong>tware Engineering Education<br />
<strong>and</strong> Training, Madrid, Spain.<br />
Prince, M., <strong>and</strong> Felder, R. (2007). "<strong>The</strong> Many Faces <strong>of</strong> Inductive Teaching <strong>and</strong><br />
Learning." Journal <strong>of</strong> College Science Teaching, 36(5), 14-20.<br />
ProjectManagementInstitute. (2008). "A Guide to the Project Management Body <strong>of</strong><br />
Knowledge (PMBOK Guide)". City: IEEE Society.<br />
Quan-Haase, A., Cothrel, J., <strong>and</strong> Wellman, B. (2005). "Instant Messaging for<br />
Collaboration: A Case Study <strong>of</strong> a High-Tech Firm." Journal <strong>of</strong> computermediated<br />
communication, 10(4).<br />
Radziwill, N. M., <strong>and</strong> Shelton, A. L. "TWiki as a Platform for Collaborative S<strong>of</strong>tware<br />
Development Management." Presented at Advanced S<strong>of</strong>tware, Control, <strong>and</strong><br />
Communication Systems for Astronomy, Glasgow, Scotl<strong>and</strong>, United Kingdom.<br />
R<strong>and</strong>all, D., Harper, R., <strong>and</strong> Rouncefield, M. (2007). Field<strong>work</strong> for Design. <strong>The</strong>ory <strong>and</strong><br />
Practice: Springer.<br />
Riel, M., <strong>and</strong> Polin, L. (2004). "Online Learning Communities. Common Ground <strong>and</strong><br />
Critical Differences in Designing Technical Environments", in S. A. K. Barab,<br />
Rob; Gray, James H., (ed.), Designing for Virtual Communities in the Service <strong>of</strong><br />
Learning. Cambridge: Cambridge University Press, pp. 16-50.<br />
Rogers, Y., <strong>and</strong> Ellis, J. (1994). "Distributed Cognition: an alternative frame<strong>work</strong> for<br />
analyzing <strong>and</strong> explaining collaborative <strong>work</strong>ing." Journal <strong>of</strong> Information<br />
Technology, 9, 119-128.<br />
Rooij, S. W. v. (2009). "Scaffolding project-based <strong>learning</strong> with the project<br />
management body <strong>of</strong> knowledge (PMBOK)." <strong>Computer</strong>s & Education, 52, 10.<br />
Sacks, H. (1992). Lectures on Conversation, Oxford, UK: Blackwell.<br />
Salomon, G. (1993). Distributed Cognitions, New York: Cambridge University Press.<br />
Schindler, M., <strong>and</strong> Eppler, M. J. (2003). "Harvesting project knowledge: a review <strong>of</strong><br />
project <strong>learning</strong> methods <strong>and</strong> success factors." International Journal <strong>of</strong> Project<br />
Management, 21, 10.<br />
Schmidt, K., <strong>and</strong> Simone, C. (1996). "Coordination Mechanisms: Towards a Conceptual<br />
Foundation <strong>of</strong> CSCW Systems Design." <strong>Computer</strong> Supported Cooperative<br />
Work, 5, 155-200.<br />
Schön, D. (1983). <strong>The</strong> Reflective Practitioner: Basic Books, Inc.<br />
Schön, D. (1987). Educating the Reflective Practitioner. , San Fransisco: Jossey-Bass.<br />
Sharp, H., <strong>and</strong> Robinson, H. (2006). "A Distributed Cognition Account <strong>of</strong> Mature XP<br />
Teams", in M. Marchesi <strong>and</strong> G. Succi, (eds.), XP 2006. Berlin Heidelberg:<br />
Springer-Verlag, pp. 1-10.<br />
Sommerville, I. (2001). S<strong>of</strong>tware Engineering, Sixth Edition: Addison-Wesley.<br />
Stafford, T. (2008). "Introduction to the Special Issue on Instant Messaging in the<br />
Workplace." IEEE Transactions on Pr<strong>of</strong>essional Communication, 51(4), 350-<br />
351.<br />
88
References<br />
Stahl, G. (2002). "Building collaborative knowing", in J.-W. Strijbos, P. A. Kirschner,<br />
<strong>and</strong> R. L. Martens, (eds.), What We Know About CSCL And Implementing It In<br />
Higher Education. Boston: Kluwer Academic Publishers, pp. 53-85.<br />
Stahl, G., Koschmann, T., <strong>and</strong> Suthers, D. (2006). "<strong>Computer</strong>-supported collaborative<br />
<strong>learning</strong>: An historical perspective", in R. K. Sawyer, (ed.), Cambridge<br />
h<strong>and</strong>book <strong>of</strong> the <strong>learning</strong> sciences. Cambridge, UK: Cambridge University<br />
Press.<br />
Star, S. L. (1990). "<strong>The</strong> Structure <strong>of</strong> Ill-Structured Solutions: Boundary Objects <strong>and</strong><br />
Heterogeneous Distributed Problem Solving", in M. Huhns <strong>and</strong> L. Gasser,<br />
(eds.), Distributed Artificial Intelligence. Morgan Kaufmann Publishers, pp. 37-<br />
54.<br />
Star, S. L., <strong>and</strong> Griesemer, J. R. (1989). "Institutional Ecology, 'Translations' <strong>and</strong><br />
Boundary Objects: Amateurs <strong>and</strong> Pr<strong>of</strong>essionals in Berkley's Museum <strong>of</strong><br />
Vertebrate Zoology, 1907-39." Social Studies <strong>of</strong> Science, 19, 387-420.<br />
Storey, M.-A. D., Cubranic, D., <strong>and</strong> German, D. M. "On the use <strong>of</strong> visualization to<br />
support awareness <strong>of</strong> human activities in s<strong>of</strong>tware development: a survey <strong>and</strong><br />
frame<strong>work</strong>." Presented at 2005 ACM symposium on S<strong>of</strong>tware visualization,<br />
St.Louis, Missouri, USA.<br />
Strauss, A. (1993). Continual permutations <strong>of</strong> action, New York: Aldine de Gruyter.<br />
Strijbos, J.-W., Kirschner, P. A., <strong>and</strong> Martens, R. L. (2004). What We Know About<br />
CSCL And Implementing It In Higher Education: Kluwer Academic Publishers.<br />
Suthers, D., <strong>and</strong> Hundhausen, C. "Learning by Constructing Collaborative<br />
Representations: An Empirical Comparison <strong>of</strong> Three Alternatives." Presented at<br />
European Conference on <strong>Computer</strong>-Supported Collaborative Learning,<br />
Maastrict, the Netherl<strong>and</strong>s.<br />
Suthers, D. D. (2006). "Technology affordances for intersubjective meaning making: A<br />
research agenda for CSCL." <strong>Computer</strong>-Supported Collaborative Learning, 1,<br />
315-337.<br />
Thomas, J. (2000). A review <strong>of</strong> research on project-based <strong>learning</strong>, Novato, CA: <strong>The</strong><br />
Buck Institute for Education.<br />
Trentin, G. (2009). "Using a Wiki to Evaluate Individual Contribution to a<br />
Collaborative Learning Project." Journal <strong>of</strong> <strong>Computer</strong> Assisted Learning, 25(1),<br />
43-55.<br />
von Glasersfeld, E. (1989). "Cognition, Construction <strong>of</strong> Knowledge, <strong>and</strong> Teaching."<br />
Synthese, 80(1), 121-140.<br />
Vygotsky, L. S. (1978). Mind in Society. <strong>The</strong> Development <strong>of</strong> Higher Psychological<br />
Processes., Cambridge, Massachusetts: Harvard University Press.<br />
Walsham, G. (1995). "Interpretive case studies in IS research: nature <strong>and</strong> method."<br />
European Journal <strong>of</strong> Information Systems(4), 74-81.<br />
Walsham, G. (2001). "Knowledge Management: <strong>The</strong> benefits <strong>and</strong> limitations <strong>of</strong><br />
computer systems." European Management Journal, 19(6), 599-608.<br />
Walsham, G. (2006). "Doing interpretive research." European Journal <strong>of</strong> Information<br />
Systems, 15, 11.<br />
Wenger, E. (1998). Communities <strong>of</strong> Practice. Learning, Meaning, <strong>and</strong> Identity:<br />
Cambridge University Press.<br />
Wenger, E. (2000). "Communities <strong>of</strong> practice <strong>and</strong> social <strong>learning</strong> systems."<br />
Organization, 7(2), 225-246.<br />
89
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Wood, D., Bruner, J., <strong>and</strong> Ross, G. (1976). "<strong>The</strong> role <strong>of</strong> tutoring in problem solving "<br />
Journal <strong>of</strong> Child Psychology <strong>and</strong> Psychiatry, 17(2), 89-100.<br />
Xiao, L., Clark, S., Rosson, M. B., <strong>and</strong> Carroll, J. M. "Promoting Reflective Thinking in<br />
Collaborative Learning Activities." Presented at Eighth IEEE International<br />
Conference on Advanced Learning Technologies (ICALT), Sant<strong>and</strong>er,<br />
Cantrabria, Spain.<br />
Xu, L. (2007). "Project the wiki way: Using wiki for computer science course project<br />
management." Journal <strong>of</strong> Computing Sciences in Colleges 22(6).<br />
Yin, R. K. (2003). Case Study Research. Design <strong>and</strong> Methods. Third Edition.: SAGE<br />
Publications.<br />
Zuhrieh, S. (2009). "Learning with Technology: Using Discussion Forums to Augment<br />
a Traditional-Style Class." Educational Technology <strong>and</strong> Society, 12(3), 15.<br />
90
Glossary<br />
B<br />
Boundary object – artifacts, documents, terms, concepts etc. around which CoPs can<br />
organize their interconnections (Wenger 1998). Boundary Objects are plastic enough<br />
to adapt to local needs <strong>and</strong> constraints <strong>of</strong> the parties employing them, yet robust<br />
enough to maintain a common identity across sites (Star <strong>and</strong> Griesemer 1989).<br />
Brokering – connections provided by people who can introduce elements <strong>of</strong> one<br />
practice into another (Wenger 1998)<br />
C<br />
Collaboration - to <strong>work</strong> jointly with others or together especially in an intellectual<br />
endeavour (Meriam Webster Online). In the literature within the research fields <strong>of</strong><br />
CSCW <strong>and</strong> CSCL, there are definitions distinguishing between collaboration <strong>and</strong><br />
cooperation, some pointing to cooperation as involving the coordination <strong>of</strong><br />
independent tasks, whereas collaboration involves mutually dependent tasks.<br />
However, the usage <strong>of</strong> these terms varies. In the thesis I do not apply a strict<br />
distinction, <strong>and</strong> in the different research papers, I have been using the terms<br />
cooperation technology, collaboration tools <strong>and</strong> collaborative tools interchangeably.<br />
<strong>Computer</strong> supported cooperative <strong>work</strong> (CSCW) – the research field addressing<br />
how collaborative activities <strong>and</strong> their coordination can be supported by means <strong>of</strong><br />
computer systems (Carstensen <strong>and</strong> Schmidt 2002 (1999)).<br />
<strong>Computer</strong> supported collaborative <strong>learning</strong> (CSCL) – an emerging branch <strong>of</strong> the<br />
<strong>learning</strong> sciences concerned with studying how people can learn together with the<br />
help <strong>of</strong> computers (Stahl et al. 2006). See Technology-Enhanced Learning (TEL).<br />
Community <strong>of</strong> practice (CoP) – A group <strong>of</strong> people characterized by a joint<br />
enterprise, mutual engagement, <strong>and</strong> a shared repertoire. Another perspective on a CoP<br />
is to consider it a shared history <strong>of</strong> <strong>learning</strong>. (Lave <strong>and</strong> Wenger 1991; Wenger 1998)<br />
Cooperation – to act or <strong>work</strong> with another or others (Merriam Webster Online). See<br />
Collaboration.<br />
91
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
L<br />
Lightweight collaboration tools – Collaboration tools that can be acquired <strong>and</strong> taken<br />
into use at low cost (e.g. money, time to learn) for individuals <strong>and</strong> their organization.<br />
A lightweight collaboration tool typically provides a limited set <strong>of</strong> features to support<br />
one aspect <strong>of</strong> collaborative <strong>work</strong> <strong>and</strong> may thus be relatively easily integrated into<br />
existing <strong>work</strong> processes (rather than imposing a certain process on the user). Many <strong>of</strong><br />
the lightweight collaboration tools are associated with Web 2.0.<br />
O<br />
Open source s<strong>of</strong>tware (OSS) – computer s<strong>of</strong>tware for which the source code <strong>and</strong><br />
certain other rights are provided under a s<strong>of</strong>tware license that permits users to study,<br />
change, <strong>and</strong> improve the s<strong>of</strong>tware. <strong>The</strong> Open Source Initiative outlines a set <strong>of</strong> criteria<br />
for an open source licence (OpenSourceInitiative 2010)<br />
P<br />
Post-mortem – In the context <strong>of</strong> project <strong>work</strong>, post-mortem refers to something<br />
undertaken after project completion or after the completion <strong>of</strong> a project phase.<br />
Problem based <strong>learning</strong> – see Section 3.1<br />
Project based <strong>learning</strong> (PBL) – see Section 3.1<br />
Project retrospective – an activity conducted after the completion <strong>of</strong> an entire project<br />
or a major project phase to reflect on <strong>and</strong> learn from the project process in order to<br />
improve it (Dingsøyr 2005) (or identify lessons learned to draw on in other projects).<br />
Post-mortem review, post-mortem analysis, post-mortem evaluation, <strong>and</strong> also, for<br />
short, post-mortem largely refer to the same type <strong>of</strong> activity, project retrospective<br />
typically being used about post-mortem evaluation in the context <strong>of</strong> agile s<strong>of</strong>tware<br />
development (Derby et al. 2006; Kerth 2001).<br />
S<br />
Social s<strong>of</strong>tware – a range <strong>of</strong> s<strong>of</strong>tware systems used for interacting <strong>and</strong> sharing data,<br />
comprising lightweight collaboration tools like instant messaging as well as web sites<br />
like Facebook, LinkedIn, Flickr, YouTube, amazon <strong>and</strong> eBay allowing the users to<br />
interact through the system <strong>and</strong> build <strong>and</strong> maintain a social pr<strong>of</strong>ile. See Web 2.0.<br />
S<strong>of</strong>tware development (SD) – the domain <strong>of</strong> <strong>work</strong> concerned with the development<br />
<strong>of</strong> s<strong>of</strong>tware. Can be considered a subset <strong>of</strong> s<strong>of</strong>tware engineering.<br />
S<strong>of</strong>tware engineering (SE) – A) An engineering discipline which is concerned with<br />
all aspects <strong>of</strong> s<strong>of</strong>tware production from early stages <strong>of</strong> system specification through to<br />
92
Glossary<br />
maintaining the system after it has been put into use (Sommerville 2001) – B) (1) <strong>The</strong><br />
application <strong>of</strong> a systematic, disciplined, quantifiable approach to the development,<br />
operation, <strong>and</strong> maintenance <strong>of</strong> s<strong>of</strong>tware; that is, the application <strong>of</strong> engineering to<br />
s<strong>of</strong>tware. (2) <strong>The</strong> study <strong>of</strong> approaches as in (1) (IEEE 1990).<br />
T<br />
Technology enhanced <strong>learning</strong> (TEL) – the support <strong>of</strong> any <strong>learning</strong> activity through<br />
technology. Whereas CSCL (see above) can be considered a distinct research<br />
community, it is also possible to consider it part <strong>of</strong> TEL. This has been done for<br />
simplicity in the thesis, in which CSCL is referred to in cases where it is relevant to<br />
point to that particular research community or body <strong>of</strong> literature.<br />
Timeline – 1) In the <strong>reflection</strong> <strong>work</strong>shops described in the thesis, a T. is a line drawn<br />
on paper or whiteboard along which are marked events perceived by project<br />
participants to be <strong>of</strong> some importance to the project. 2) A collaboration tool (e.g.<br />
Trac) may include functionality for chronologically displaying (<strong>of</strong>ten fine grained)<br />
actions/events related to the <strong>work</strong> supported by the tool (e.g. all source code updates).<br />
In Trac, the display <strong>of</strong> this chronology is named the Timeline.<br />
Trajectory - “(1) the course <strong>of</strong> any experienced phenomenon as it evolves over time”<br />
<strong>and</strong> “(2) the actions <strong>and</strong> interactions contributing to its evolution” (Strauss 1993,<br />
pp.53-54).<br />
U<br />
Use – the act or practice <strong>of</strong> employing something (Webster Online).<br />
Usage – firmly established <strong>and</strong> generally accepted practice or procedure (Webster<br />
Online).<br />
W<br />
Web 2.0 – This is a concept for which many different definitions <strong>and</strong> explanations<br />
exist. Dohn describes Web 2.0 as practice rather than a set <strong>of</strong> technologies, arguing<br />
that a particular technology may be used in a way that is more or less Web 2.0. Web<br />
2.0, according to Dohn, denotes activities that have most or all <strong>of</strong> the following<br />
characteristics (the last one being necessary):<br />
• “Collaboration <strong>and</strong>/or distributed authorship<br />
• Active, open-access, “bottom-up” participation <strong>and</strong> interactive multi-way<br />
communication<br />
93
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
• Continuous production, reproduction, <strong>and</strong> transformation <strong>of</strong> material in use<br />
<strong>and</strong> reuse across contexts<br />
• Openness <strong>of</strong> content, renunciation <strong>of</strong> copyright, distributed ownership<br />
• Lack <strong>of</strong> finality, “awareness-in-practice” <strong>of</strong> the “open-endedness” <strong>of</strong> the<br />
activity<br />
• Taking place on the WWW, or to a large extent utilizing Web-mediated<br />
resources <strong>and</strong> activities” (Dohn 2009, p.345)<br />
94
Appendix A: Research papers<br />
P1<br />
P2<br />
P3<br />
Krogstie, B. <strong>and</strong> B. Bygstad. Cross-Community Collaboration <strong>and</strong> Learning in<br />
Customer-Driven S<strong>of</strong>tware Engineering Student Projects. Twentieth<br />
Conference on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training (CSEE&T)<br />
2007, Dublin. IEEE <strong>Computer</strong> Society. (Krogstie <strong>and</strong> Bygstad 2007)<br />
Krogstie, B. Power through brokering. OSS participation in SE projects. ICSE<br />
2008, Leipzig. IEEE <strong>Computer</strong> Society. (Krogstie 2008a)<br />
Krogstie, B.R. Do‟s <strong>and</strong> Don‟ts <strong>of</strong> Instant Messaging in Students‟ Project<br />
Work. NOKOBIT 2009, Trondheim. Tapir. (Krogstie 2009a)<br />
P4 Krogstie, B.R. <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. in COOP 2008.<br />
Carry-le-Rouet, Provence, France. Institut d‟Etudes Politiques d‟Aix-en-<br />
Provence. (Krogstie 2008b)<br />
P5<br />
P6<br />
P7<br />
P8<br />
Krogstie, B.R. Using Project Wiki History to Reflect on the Project Process.<br />
42nd Hawaii International Conference on System Sciences (HICSS‟42) 2009.<br />
Big Isl<strong>and</strong>, Hawaii: IEEE <strong>Computer</strong> Society. (Krogstie 2009c)<br />
Krogstie, B.R. <strong>and</strong> M. Divitini. Shared timeline <strong>and</strong> individual experience:<br />
Supporting retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering teams.<br />
CSEE&T 2009, Hyderabad. IEEE <strong>Computer</strong> Society. (Krogstie <strong>and</strong> Divitini<br />
2009)<br />
Krogstie, B.R. A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong><br />
utilizing historical data in collaborative tools. EC-TEL 2009, Nice. Springer.<br />
(Krogstie 2009b)<br />
Krogstie, B.R. <strong>and</strong> M. Divitini. Supporting Reflection in S<strong>of</strong>tware<br />
Development with Everyday Working Tools. COOP 2010, Aix-en-Provence,<br />
France. Springer. (Krogstie <strong>and</strong> Divitini 2010)<br />
95
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
96
Research paper P1<br />
Title:<br />
Cross-Community Collaboration <strong>and</strong> Learning in Customer-Driven S<strong>of</strong>tware<br />
Engineering Student Projects<br />
Authors:<br />
Birgit Krogstie <strong>and</strong> Bendik Bygstad<br />
Published in:<br />
Proceedings <strong>of</strong> CSEE&T 2007<br />
Pages:<br />
8<br />
Complete reference:<br />
Krogstie, B. <strong>and</strong> B. Bygstad (2007). Cross-Community Collaboration <strong>and</strong> Learning in<br />
Customer-Driven S<strong>of</strong>tware Engineering Student Projects. Twentieth Conference on<br />
S<strong>of</strong>tware Engineering Education <strong>and</strong> Training (CSEE&T), Dublin, Irel<strong>and</strong>, 3-5 July.<br />
IEEE <strong>Computer</strong> Society.<br />
97
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
98
Cross-Community Collaboration <strong>and</strong> Learning in Customer-Driven<br />
S<strong>of</strong>tware Engineering Student Projects<br />
Birgit Krogstie<br />
NTNU<br />
birgitkr@idi.ntnu.no<br />
Bendik Bygstad<br />
NITH<br />
bendik.bygstad@nith.no<br />
Abstract<br />
This paper explores collaboration <strong>and</strong> <strong>learning</strong> between stakeholders in customer-driven<br />
student projects. <strong>The</strong> research objectives are to obtain empirically based knowledge on how<br />
students relate to stakeholders in customer-driven projects, <strong>and</strong> to suggest implications for<br />
the pedagogical design <strong>of</strong> the project courses.<br />
Empirical data was collected from two Bachelor courses in s<strong>of</strong>tware engineering at two<br />
<strong>learning</strong> institutions in Norway. To make sense <strong>of</strong> the interaction between the three<br />
stakeholders in the project: the student groups, the university <strong>and</strong> the customer, we build on<br />
Wenger’s concept <strong>of</strong> communities <strong>of</strong> practice <strong>and</strong> on the concept <strong>of</strong> boundary objects.<br />
Our analysis highlights that students, through practical experience in the projects, learn to<br />
balance the requirements <strong>and</strong> expectations from different stakeholders in designing a <strong>work</strong>ing<br />
technical solution - a valuable skill for s<strong>of</strong>tware engineers. We argue that for students to<br />
learn to balance stakeholders’ interests in the best possible way, visibility <strong>of</strong> stakeholders’<br />
goals should be focused throughout the projects. Explicit reference to the goals should be<br />
incorporated into project artifacts serving as boundary objects. Collaboration technologies<br />
providing st<strong>and</strong>ard shared <strong>work</strong>space functionality are seen as adequate to support this.<br />
1. Introduction<br />
S<strong>of</strong>tware Engineering (SE) project courses provide students with arenas for project<br />
based <strong>learning</strong> [1, 2] <strong>and</strong> serve as a stage <strong>of</strong> transition from the student role to that <strong>of</strong> a<br />
SE pr<strong>of</strong>essional. Projects with external customers <strong>and</strong> “real” problems provide<br />
authenticity [3] <strong>and</strong> have been found to be useful in preparing the students for <strong>work</strong> in<br />
the IT industry [4]. A pedagogical ideal for the projects can be seen to have the students<br />
complete a tour <strong>of</strong> “educational refinement” in the SE “real world”, making use <strong>of</strong> their<br />
knowledge <strong>and</strong> skills to their best <strong>of</strong> their ability while having their beliefs challenged.<br />
Through the projects, we want students to make a first step towards practicing the SE<br />
‘methodology-in-action’ [5] which requires enough experience to allow for a critical<br />
distance to, <strong>and</strong> application <strong>of</strong>, the formalized SE methodology taught at the university.<br />
In SE projects, success depends on the team recognizing <strong>and</strong> reconciling the different<br />
goals, knowledge <strong>and</strong> practices among major stakeholders [6, 7]. In customer-driven<br />
student projects, there is one stakeholder not found in pr<strong>of</strong>essional SE: the university as<br />
represented by the course staff. <strong>The</strong> latter serves as a resource assisting the students in<br />
meeting realistic challenges <strong>of</strong> SE <strong>work</strong>, e.g. managing the customer relationship, but at<br />
the same time, they impose complexity <strong>and</strong> constraints that are not there in real SE<br />
<strong>work</strong>. <strong>The</strong> research objectives <strong>of</strong> the <strong>work</strong> presented in this paper are to obtain<br />
empirically based knowledge on how students relate to stakeholders in customer-driven<br />
projects, <strong>and</strong> to suggest implications for the pedagogical design <strong>of</strong> the project courses.<br />
In what follows, we first introduce our theoretical underst<strong>and</strong>ing <strong>of</strong> stakeholder<br />
communities <strong>and</strong> goals, <strong>and</strong> how students’ reconciliation <strong>of</strong> them is essential to<br />
<strong>learning</strong>. Particularly, we point to the role <strong>of</strong> artifacts as boundary objects between<br />
communities. In section 3 we present our case study. In section 4 we provide a number<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
99
<strong>of</strong> examples from the case illustrating how students reflect on their relationship to<br />
different communities in the project. In section 5, we discuss our findings in the light<br />
<strong>of</strong> theory <strong>and</strong> suggest implications for the pedagogical design <strong>of</strong> project courses.<br />
Finally, we address limitations <strong>and</strong> further <strong>work</strong>, <strong>and</strong> conclude the paper.<br />
2. Cross-community collaboration: the role <strong>of</strong> goals <strong>and</strong> boundary objects<br />
In this section we elaborate on three generic aspects <strong>of</strong> customer-driven SE student<br />
projects that can be seen as central to students’ <strong>learning</strong>: stakeholders as <strong>learning</strong><br />
communities, stakeholders’ goals for project engagement, <strong>and</strong> project activity as a<br />
question <strong>of</strong> relating to <strong>and</strong> reconciling the goals <strong>and</strong> interests <strong>of</strong> interrelated <strong>learning</strong><br />
communities.<br />
2.1. Project stakeholders represent different kinds <strong>of</strong> <strong>learning</strong> communities<br />
Learning communities are communities designed to support <strong>learning</strong>. Riel <strong>and</strong> Polin<br />
[8] describe three, overlapping types: Task-based <strong>learning</strong> communities are groups <strong>of</strong><br />
people who are organized around a task, <strong>work</strong>ing together for a specified period <strong>of</strong> time<br />
to produce a product. Practice-based <strong>learning</strong> communities are larger groups with<br />
shared goals <strong>and</strong> provide members with richly contextualized <strong>and</strong> supported arenas for<br />
<strong>learning</strong>. Knowledge-based <strong>learning</strong> communities resemble the practice-based ones but<br />
are focused on producing external knowledge about the practice.<br />
A university is <strong>of</strong>ficially designated to serve a role in society as a practice-based <strong>and</strong><br />
knowledge-based <strong>learning</strong> community. Its practices include research <strong>and</strong><br />
teaching/<strong>learning</strong>. A part <strong>of</strong> the university organization, i.e. a SE study program, can be<br />
considered a <strong>learning</strong> community <strong>of</strong> its own. A customer organization may also be<br />
considered both practice-based <strong>and</strong> knowledge-based, the weight on the latter<br />
depending on the actual policies <strong>and</strong> practices (e.g. knowledge management efforts). A<br />
sub-community in the customer organization may typically be serving as the <strong>learning</strong><br />
community most directly relating to the student project. <strong>The</strong> customer-driven SE<br />
student project exists for a much more limited time than the other two communities, but<br />
it can be considered as a type <strong>of</strong> <strong>learning</strong> community focused on the achievement <strong>of</strong> a<br />
task: a task-based <strong>learning</strong> community.<br />
University<br />
SE study program<br />
SE pr<strong>of</strong>ession<br />
<strong>learning</strong><br />
Customer<br />
organization<br />
SE student<br />
project<br />
Figure 1: <strong>The</strong> SE project at the intersection <strong>of</strong> <strong>learning</strong> communities,<br />
Several overlapping communities will potentially learn from the project (see Figure 1). <strong>The</strong><br />
impact is likely to be smaller, <strong>and</strong> the <strong>learning</strong> <strong>cycle</strong> longer, the more general the community.<br />
E.g., the project supervisor may learn about project supervision from one session to the next;<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
100
the course staff may improve the course between semesters, <strong>and</strong> the success <strong>of</strong> the SE study<br />
program may finally influence university decisions. <strong>The</strong> customer representative may learn<br />
about technical issues whereas her department benefits from knowledge about the viability <strong>of</strong><br />
an idea after project completion.<br />
2.2 Project stakeholders have different goals for their engagement in the projects<br />
Students want formal qualifications making them attractive on the job market, which<br />
includes knowledge <strong>and</strong> skill e.g. in project management <strong>and</strong> programming. <strong>The</strong> project<br />
customer becomes part <strong>of</strong> the students’ net<strong>work</strong> in the business world, perhaps leading<br />
to a job <strong>of</strong>fer. Students also want to enjoy the social aspects <strong>of</strong> project <strong>work</strong>.<br />
<strong>The</strong> course staff seeks to improve the course. Such improvement partially rests on<br />
evaluation <strong>of</strong> the projects as reflected in the grades. Learning objectives are formalized<br />
in course descriptions. Students’ viewpoints on the projects (e.g. as found in <strong>reflection</strong><br />
notes) provide feedback on the pedagogical quality <strong>of</strong> the project course. Additionally,<br />
the project course must meet relevant needs in the business world. If customers express<br />
that their needs are met, it is an acknowledgement <strong>of</strong> the relevance <strong>of</strong> the project course<br />
<strong>and</strong> courses on which the project is based. SE projects thus serve as a test bench for the<br />
SE study program. <strong>The</strong> university-customer relationship further contributes in building<br />
a business net<strong>work</strong> for the university, to be utilized in new student or research projects.<br />
2.3. Cross-community interaction as essential to project <strong>work</strong> <strong>and</strong> <strong>learning</strong><br />
Two important types <strong>of</strong> connections between communities are boundary objects <strong>and</strong><br />
brokering [9]. Boundary objects are “artifacts, documents, terms, concepts, <strong>and</strong> other<br />
forms <strong>of</strong> reification around which communities <strong>of</strong> practice can organize their<br />
interconnections.” [10-12]. <strong>The</strong> artifacts central to the SE project are important because<br />
they support the goal attainment <strong>of</strong> one or more stakeholders. Reaching agreement on<br />
the development <strong>of</strong> an artifact important to more than one community is a question <strong>of</strong><br />
making alignments based on the views, knowledge <strong>and</strong> practices <strong>of</strong> the communities.<br />
An example from the SE world explicitly referring to negotiation in development <strong>work</strong><br />
is the WinWin methodology [13]. In SE student projects, there are artifacts to be<br />
negotiated with the course staff, with the customer <strong>and</strong> with both <strong>of</strong> them. <strong>The</strong><br />
requirements specification is a particularly central artifact, negotiated between the<br />
student team <strong>and</strong> the customer <strong>and</strong> subject to course staff’s supervision <strong>and</strong> evaluation.<br />
For a boundary object to <strong>work</strong> well, communities should be able to use it in making<br />
sense <strong>of</strong> their own activity as well as the activity <strong>of</strong> the other community. <strong>The</strong> boundary<br />
object contributes to making visible stakeholders’ objectives, as well as possibly the<br />
negotiation <strong>and</strong> development process, e.g. by documenting arguments, decisions <strong>and</strong><br />
versions. This corresponds to the boundary object being socially translucent [14],<br />
providing awareness, accountability <strong>and</strong> visibility <strong>of</strong> the development process.<br />
Brokering is “connections provided by people who can introduce elements <strong>of</strong> one<br />
practice into another.” [9] (p.105). Students in a SE project can be considered brokers<br />
between the university, the customer organization <strong>and</strong> the project team to the extent that<br />
they are members <strong>of</strong> all communities <strong>and</strong> <strong>work</strong> to align the perspectives <strong>and</strong> practices<br />
from each community in the project. In the capacity <strong>of</strong> being students, the developers<br />
are participants in the teaching/<strong>learning</strong> practices <strong>of</strong> the university. Also, they are<br />
participants in the task-oriented practice <strong>of</strong> the project team. Further, the students can<br />
be seen as novice pr<strong>of</strong>essionals engaging in legitimate peripheral participation [15] in<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
101
the SE pr<strong>of</strong>essional community, but the team members generally have no formal<br />
membership in the project customer organization.<br />
3. <strong>The</strong> case: two project courses in Norway<br />
In an interpretive qualitative study we collected data from the spring semester <strong>of</strong><br />
2006 in two, fairly similar bachelor level project courses in Norway. <strong>The</strong> authors are<br />
engaged in teaching <strong>and</strong> supervision <strong>of</strong> these project courses. Project Course 1 (P1) is a<br />
project course at NITH, which is a small, private university college delivering study<br />
programs within IT. Project Course 2 (P2) is a course at NTNU, which is a large public<br />
university with a main focus on technical study programs.<br />
Although there are some differences between the courses, they are structurally<br />
similar enough (see Table 1) for us to use empirical data from both <strong>of</strong> them combined to<br />
analyze the relationship between the stakeholders.<br />
Table 1: Characteristics <strong>of</strong> the two project courses <strong>of</strong> our case<br />
Characteristics Project Course 1 (NITH) Project Course 2 (NTNU)<br />
Level<br />
Bachelor, 6 th (final) semester<br />
Number <strong>of</strong> students 50<br />
46<br />
Size <strong>of</strong> project Approx.5 students, self-formed Approx.5 students, mostly self-formed<br />
groups/duration groups, one semester, 60% <strong>of</strong> groups, one semester, 50% <strong>of</strong> semester<br />
semester <strong>work</strong>load<br />
<strong>work</strong>load<br />
Form <strong>of</strong> course Only introductory / occasional lectures. Supervisor from staff. Appointed group<br />
delivery<br />
contact in the customer organization.<br />
Location Work mainly at the customer’s site. Work mainly on the university campus.<br />
Some groups <strong>work</strong> at customers’ site.<br />
Educational<br />
Mostly common courses, including Some common courses, including basic<br />
background within basic courses in Java <strong>and</strong><br />
system analysis <strong>and</strong> programming.<br />
the study program RUP/UML<br />
Previous project 2 nd year, case ready-made for the 2 nd year, case ready-made for the<br />
course<br />
course, RUP-based, Java/UML, no course, partly XP-based, Java/UML, no<br />
external customer<br />
external customer<br />
Case Real case, external customer, s<strong>of</strong>tware product to be developed<br />
Imposed structure on Final project report required. Method (e.g. process model) freely chosen by the<br />
part <strong>of</strong> the school group. Groups required to report regularly on progress/status to supervisor.<br />
Mid-term report required.<br />
Assessment One grade for the whole group. Both product, process <strong>and</strong> presentation count.<br />
Data for this paper were collected through semi-structured interviews with P1<br />
groups, done as part <strong>of</strong> an exploratory case study on <strong>work</strong> <strong>and</strong> collaboration support in<br />
the projects. One-hour interviews were conducted ca 3 weeks before project delivery<br />
with 7 out <strong>of</strong> 11 groups. We also made brief (10-20 min) telephone interviews with P1<br />
<strong>and</strong> P2 customers 2-3 months after project completion, addressing customers’ rationale<br />
for, <strong>and</strong> outcome from, project participation. 6 out <strong>of</strong> 14 P1 customers responded, as did<br />
9 out <strong>of</strong> 9 P2 customers (<strong>of</strong> 10 projects). We further draw on <strong>reflection</strong> notes in<br />
students’ project reports. Also, we made extensive use <strong>of</strong> available written documents:<br />
customers’ evaluation forms, formal course specifications, project reports, etc.<br />
Main viewpoints from the customer interviews were categorized <strong>and</strong> summarized in a<br />
table. Data on the project teams was analyzed in the following steps. First, the<br />
interviews <strong>and</strong> the collected written materials were summarized for each group. <strong>The</strong>n<br />
we used our theoretical lens, described in the previous section, to identify central<br />
attributes <strong>of</strong> the experiences for each group. Finally, we analyzed the data across<br />
groups, identifying patterns <strong>and</strong> explanations.<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
102
4: Case findings: students’ view on relating to different communities<br />
Our analysis resulted in four main findings, the first one relating to customers’<br />
viewpoints on the projects <strong>and</strong> the others relating to students’ viewpoints.<br />
4.1. Customers have several different goals for their project participation<br />
Customer interviews indicated satisfaction with the project; products having been<br />
developed according to specifications <strong>and</strong> expectations <strong>and</strong> at very low cost. Many<br />
customers reported that used the project to learn about <strong>and</strong>/or evaluate their technology<br />
<strong>and</strong>/or ideas. Also, many said they had learnt something that would influence their daily<br />
<strong>work</strong>. Further, many reported <strong>learning</strong> about playing the role <strong>of</strong> customer, e.g. in terms<br />
<strong>of</strong> needs for involvement. <strong>The</strong> requirements specification was generally regarded by<br />
customers as the most important project document. Finally, many customers were<br />
looking for new employees, <strong>and</strong> some ended up recruiting from the project groups.<br />
4.2. Project teams relate to communities <strong>and</strong> community membership<br />
We find that SD project students actively relate to the various communities involved<br />
<strong>and</strong> the balancing <strong>of</strong> requirements from different project stakeholders<br />
In our interviews we asked the students how they mainly saw themselves <strong>work</strong>ing in<br />
the project (e.g. at the customer’s site): as students or consultants/pr<strong>of</strong>essionals.<br />
Students generally saw themselves as students. <strong>The</strong>re was some indication that groups<br />
who most strongly considered their <strong>work</strong> to be <strong>of</strong> real value to the customer, also<br />
considered themselves more as consultants. <strong>The</strong> customer’s expressed confidence in the<br />
group, expectations for the result <strong>and</strong> willingness to provide necessary resources<br />
seemed to boost students’ self confidence about their role as novice pr<strong>of</strong>essionals.<br />
Most groups considered their project to be close to a real SE project, with the<br />
exception <strong>of</strong> the course staff’s requirements for documentation, the lack <strong>of</strong><br />
consequences if the project failed <strong>and</strong> the general absence <strong>of</strong> economic considerations.<br />
“This is no Mickey Mouse project”, as the project manager in one group expressed it.<br />
4.3 Awareness <strong>of</strong> stakeholder goals is <strong>of</strong>ten insufficient in the teams<br />
Students are not necessarily aware <strong>of</strong> all stakeholder goals for the project <strong>work</strong> (e.g.<br />
as revealed in our customer interviews), or they have an interpretation <strong>of</strong> stakeholder<br />
goals that might have been questioned by the stakeholders themselves if they had<br />
known students’ reasoning.<br />
Students are aware that different stakeholder goals should be met, some related to<br />
completing <strong>work</strong> as a SE project <strong>and</strong> others to completing the project as a university<br />
course: “In a real project there would have been money involved. Now focus is very<br />
much on us making a project following the school’s requirements, even if some <strong>of</strong> the<br />
school’s requirements is that it should be good for the customer, but here in a way we<br />
<strong>work</strong> to have the best possible grade, apart from <strong>work</strong>ing on a product that sort <strong>of</strong> leads<br />
to further contracts; the customer is happy.”<br />
We found a general opinion among the students that the role <strong>of</strong> the university is to<br />
provide <strong>and</strong> test students’ use <strong>of</strong> “by the book” SE knowledge. <strong>The</strong>re is a largely<br />
implicit requirement that guidelines <strong>and</strong> practices (e.g. <strong>of</strong> modeling, diagramming <strong>and</strong><br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
103
programming) taught in other courses be followed. Students contrast the SE practices<br />
prescribed by the course staff to the customer’s requirements <strong>and</strong> ‘the real world’: “I do<br />
not think companies [..] need so much. [..] as what we do [..] in school. Because as it<br />
turned out, we were writing a lot <strong>of</strong> reports, <strong>and</strong> then our supervisor from <br />
said I do not need all these pages, I just want concrete things”.<br />
4.4. Potential boundary objects are not actual boundary objects in the projects<br />
Project artifacts intended by staff to be boundary objects are not always perceived by<br />
students as relevant to stakeholders’ goals. For instance, certain models <strong>and</strong> associated<br />
diagrams are typically perceived as unnecessary or purely 'school knowledge'. Students<br />
do however reflect on the difference between models supporting development <strong>work</strong> <strong>and</strong><br />
models for cross-community communication. “..[].. the models <strong>and</strong> the documentation<br />
we make is mainly to give us an underst<strong>and</strong>ing; they are not so much for them [..]. It is not necessarily the same models, it is not necessarily done in the<br />
same way.” Also, the course staff’s requirements are contrasted with the group’s<br />
needs:.”[I] have on several occasions felt that ’this model, we do not really need it’, but<br />
we have to spend time on it because the school requires it to be there.”<br />
5. Implications for the pedagogical design <strong>of</strong> customer-driven SE projects<br />
It is not problematic that customers have several goals for their project participation,<br />
as long as students’ <strong>learning</strong> is not as a consequence being downgraded in the projects.<br />
<strong>The</strong> involvement <strong>of</strong> course staff usually ensures a focus on the formal <strong>learning</strong> goals.<br />
Customers’ various goals, in our view, underscore the potential <strong>of</strong> customer-driven SE<br />
projects to serve a role beyond the confines <strong>of</strong> the SE study program.<br />
It is however a challenge to have all stakeholder goals made explicit. Our findings<br />
indicate that students’ interpretation <strong>of</strong> stakeholder goals is inadequate, making it<br />
difficult for students to decide what project artifacts are important to what stakeholders,<br />
<strong>and</strong> in what way. Customers <strong>and</strong> course staff also have an interest in knowing about the<br />
goals <strong>and</strong> students’ underst<strong>and</strong>ing <strong>of</strong> them; this allows for elaboration on the meaning<br />
<strong>of</strong> the goals <strong>and</strong> negotiation over their importance when priorities must be made.<br />
One aspect which should be kept in mind as we stress whether students give<br />
m<strong>and</strong>atory or recommended project artifacts a role as boundary objects in their <strong>work</strong>, is<br />
their purpose: effective cross-community collaboration to reach a good project result. If<br />
students find a better way <strong>of</strong> collaborating with their customer, e.g. over a different<br />
type <strong>of</strong> boundary object, course staff should give due credits to the approach. In terms<br />
<strong>of</strong> pedagogy, the project supervisor must be able to see when advice should be given to<br />
adhere to st<strong>and</strong>ard artifacts in a prescribed way, <strong>and</strong> when students should be ‘let loose’<br />
to follow their own approach. Underlying our viewpoint here is the assumption that<br />
students’ <strong>learning</strong> benefits both from their active development <strong>of</strong> their own solutions<br />
<strong>and</strong> from an experience <strong>of</strong> mastery in their project [16].<br />
A related issue is the perceived discrepancy between real life <strong>and</strong> school knowledge<br />
among the students in our study. This could be interpreted as a useful insight on part <strong>of</strong><br />
the students, but also as a lack <strong>of</strong> ability to see the applicability <strong>of</strong> particular tools <strong>and</strong><br />
techniques in a real project. To have students experience the usefulness <strong>of</strong> various<br />
project artifacts, it may be pedagogically necessary to make m<strong>and</strong>atory the development<br />
<strong>of</strong> certain artifacts. It is however essential to have students actively relate to the link<br />
between artifacts <strong>and</strong> stakeholder goals, which implies having students provide their<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
104
own account <strong>of</strong> why each artifact is <strong>of</strong> importance to one or more <strong>of</strong> the stakeholders.<br />
<strong>The</strong>is account should be based on communication with the project stakeholders.<br />
<strong>The</strong> set <strong>of</strong> connections between artifacts <strong>and</strong> goals is in itself a boundary object,<br />
typically manifest in documents such as a team/customer collaboration agreement <strong>and</strong> a<br />
S<strong>of</strong>tware Development Plan, <strong>and</strong> typically developing throughout the project.<br />
5.2. Implications for the pedagogical design <strong>of</strong> customer-driven SE project courses<br />
Building on the previous discussion, we suggest that the following guidelines be<br />
followed in the design <strong>of</strong> customer-driven project courses:<br />
Stakeholders should be encouraged to make their goals explicit to themselves <strong>and</strong><br />
each other. Students should be aided in eliciting customer requirements. Also, they<br />
should be encouraged to discuss team members’ objectives for project participation as<br />
well as the goals for the team as a whole. Developing a set <strong>of</strong> team-internal project<br />
rules may be useful. Course staff needs to make sure that <strong>learning</strong> goals <strong>and</strong> evaluation<br />
criteria are clearly defined <strong>and</strong> easily available to students <strong>and</strong> project customers.<br />
Course staff should also check on students’ underst<strong>and</strong>ing <strong>of</strong> the goals. Customers need<br />
guidance about their role <strong>and</strong> about conveying their objectives as clearly as possible to<br />
the group: ‘<strong>learning</strong> to be a customer’ should occur as early as possible in the project.<br />
For every project artifact serving a role as a boundary object in the project, its role<br />
for the attainment <strong>of</strong> project goals as well as the rationale behind its current state should<br />
be made explicit, opening up for scrutiny <strong>and</strong> negotiation among project stakeholders.<br />
A representation <strong>of</strong> the relationship between artifacts <strong>and</strong> stakeholder goals should be<br />
incorporated into the artifacts themselves. This can be achieved by imposing the use <strong>of</strong><br />
st<strong>and</strong>ards <strong>and</strong> templates explicitly referring to stakeholder goals. Also, there are<br />
st<strong>and</strong>ard functionality shared <strong>work</strong>space tools adequate for providing the necessary<br />
persistence, content structuring, <strong>work</strong>space awareness <strong>and</strong> access for the communities<br />
involved. Such tools should be utilized in the projects.<br />
5.3. Limitations to our study<br />
Our case material is limited in scope <strong>and</strong> time, although we draw on materials from<br />
two different <strong>learning</strong> institutions. In-depth interviews with all relevant stakeholders,<br />
combined with a rich written material should lead to satisfactory internal validity <strong>of</strong><br />
findings. External validity is mainly determined by our research approach. An<br />
intensive, qualitative approach tends to produce results strong on accuracy, but weaker<br />
on generalization [17]. This indicates that our findings will be most relevant in contexts<br />
similar to our cases, i.e. for customer driven projects in SE at the Bachelor level.<br />
6. Conclusion<br />
Our empirical findings indicate that students consciously balance stakeholder goals<br />
in their projects, but that the underst<strong>and</strong>ing <strong>of</strong> these goals <strong>and</strong> their connection to<br />
various project artifacts <strong>of</strong>ten is inadequate. We suggest that in the pedagogical design<br />
<strong>of</strong> the courses, a stronger focus should be placed on the connections between<br />
stakeholder goals <strong>and</strong> project artifacts. <strong>The</strong>se connections should be made visible <strong>and</strong><br />
negotiable in connection with boundary objects: artifacts playing a role in the goal<br />
attainment <strong>of</strong> more than one stakeholder Scaffolding for effective development <strong>of</strong> these<br />
artifacts implies a focus on project process as well as on technological infrastructure.<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
105
In our project courses we currently see the successful use <strong>of</strong> various shared<br />
<strong>work</strong>space tools: wikis, forums, simple file hierarchies, <strong>and</strong> more comprehensive tools<br />
(e.g. Sharepoint), used for within-team purposes <strong>and</strong> for sharing contents <strong>and</strong><br />
interacting with other stakeholders. We suggest that this infrastructure be utilized in<br />
efforts to provide effective representations <strong>of</strong>, access to, boundary objects. In future <strong>and</strong><br />
ongoing <strong>work</strong>, we investigate more closely the potential <strong>of</strong> particular collaboration<br />
tools to support the development <strong>of</strong> boundary objects in customer-driven SD projects.<br />
7. Acknowledgements<br />
[13]<br />
We wish to thank Glenn Munkvold for useful discussion related to this paper.<br />
8. References<br />
[1] P. C. Blumenfeld, E. Soloway, R. W. Marx, J. S. Krajcik, M. Guzdial, <strong>and</strong> A. Palinscar, "Motivating<br />
Project-Based Learning: Sustaining the Doing, Supporting the Learning," Educational Psychologist, vol.<br />
26, pp. 369-398, 1991.<br />
[2] L. Helle, P. Tynjälä, <strong>and</strong> E. Olkinuora, "Project-based <strong>learning</strong> in post-secondary education - theory,<br />
practice <strong>and</strong> rubber sling shots," Higher Education, vol. 51, pp. 287-314, 2006.<br />
[3] J. S. Brown, A. Collins, <strong>and</strong> P. Duguid, "Situated Cognition <strong>and</strong> the Culture <strong>of</strong> Learning," Educational<br />
Researcher, vol. 18, pp. 32-42, 1989.<br />
[4] P. J. A. van Vliet <strong>and</strong> L. R. Pietron, "Information Systems Development Education in the Real World -<br />
A Project Methodology <strong>and</strong> Assessment," Journal <strong>of</strong> Information Systems Education, vol. 17, pp. 285-<br />
293, 2006.<br />
[5] B. Fitzgerald, "<strong>The</strong> use <strong>of</strong> systems development methodologies in practice: a field study," Information<br />
Systems Journal, vol. 7, pp. 201-212, 1997.<br />
[6] J.-H. Ahn <strong>and</strong> A. E. Skudlark, "Resolving conflict <strong>of</strong> interest in the process <strong>of</strong> an information system<br />
implementation for advanced telecommunication services," Journal <strong>of</strong> Information Technology, vol. 12,<br />
pp. 3-13, 1997.<br />
[7] A. Pouloudi, "Aspects <strong>of</strong> the Stakeholder Concept <strong>and</strong> their Implications for Information Systems<br />
Development," presented at 32nd Hawaii International Conference on Systems Sciences, 1999.<br />
[8] M. Riel <strong>and</strong> L. Polin, "Online Learning Communities. Common Ground <strong>and</strong> Critical Differences in<br />
Designing Technical Environments," in Designing for Virtual Communities in the Service <strong>of</strong> Learning,<br />
S. A. K. Barab, Rob; Gray, James H., Ed. Cambridge: Cambridge University Press, 2004, pp. 16-50.<br />
[9] E. Wenger, Communities <strong>of</strong> Practice. Learning, Meaning, <strong>and</strong> Identity: Cambridge University Press,<br />
1998.<br />
[10] U. Gal, Y. Youngjin, <strong>and</strong> R. Bol<strong>and</strong>, "<strong>The</strong> dynamics <strong>of</strong> boundary objects, social infrastructures <strong>and</strong><br />
social identities," 2005.<br />
[11] H. Karsten, K. Lyytinen, M. Hurskainen, <strong>and</strong> T. Koskelainen, "Crossing boundaries <strong>and</strong> conscripting<br />
participation: representing <strong>and</strong> integrating knowledge in a paper machinery project," European Journal<br />
<strong>of</strong> Information Systems, vol. 10, pp. 89-98, 2001.<br />
[12] S. L. Star, "<strong>The</strong> Structure <strong>of</strong> Ill-Structured Solutions: Boundary Objects <strong>and</strong> Heterogeneous Distributed<br />
Problem Solving," in Distributed Artificial Intelligence, vol. 3, M. Huhns <strong>and</strong> L. Gasser, Eds.: Morgan<br />
Kaufmann Publishers, 1990, pp. 37-54.<br />
H. In, B. Boehm, T. Rodgers, <strong>and</strong> M. Deutsch, "Applying WinWin to quality requirements: a case<br />
study," presented at ICSE'01, Toronto, Ontario, Canada, 2001.<br />
[14] T. Erickson <strong>and</strong> W. A. Kellogg, "Social Translucence: An Approach to Designing Systems that Support<br />
Social Processes," ACM Transactions on <strong>Computer</strong>-Human Interaction, vol. 7, pp. 59-83, 2000.<br />
[15] J. Lave <strong>and</strong> E. Wenger, Situated Learning. Legitimate peripheral participation. Cambridge: University<br />
<strong>of</strong> Cambridge Press., 1991.<br />
[16] A. B<strong>and</strong>ura, "Self-efficacy," in Encyclopedia <strong>of</strong> human behaviour, vol. 4, V. S. Ramachaudran, Ed. New<br />
York, USA: Academic Press, 1994, pp. 71-81.<br />
[17] A. Langley, "Strategies for theorizing from process data," Academy <strong>of</strong> Management Review, vol. 24, pp.<br />
691-710, 1999.<br />
20th Conference on S<strong>of</strong>tware Engineering Education & Training (CSEET'07)<br />
0-7695-2893-7/07 $20.00 © 2007<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 08:55 from IEEE Xplore. Restrictions apply.<br />
106
Research paper P2<br />
Title:<br />
Power Through Brokering: Open Source Community Participation in S<strong>of</strong>tware<br />
Engineering Student Proejcts<br />
Authors:<br />
Birgit Krogstie<br />
Published in:<br />
Proceedings <strong>of</strong> ICSE 2008<br />
Pages:<br />
10<br />
Complete reference:<br />
Krogstie, B. (2008). Power through brokering. OSS participation in SE projects.<br />
International Conference on S<strong>of</strong>tware Engineering (ICSE) 2008, Leipzig, Germany,<br />
10-18 May. IEEE <strong>Computer</strong> Society.<br />
107
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
108
Power Through Brokering: Open Source Community<br />
Participation in S<strong>of</strong>tware Engineering Student Projects<br />
Birgit R. Krogstie<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
Sem Sæl<strong>and</strong>s vei 5-7, 7034 Trondheim, Norway<br />
birgitkr@idi.ntnu.no<br />
ABSTRACT<br />
Many s<strong>of</strong>tware engineering projects use open source s<strong>of</strong>tware tools<br />
or components. <strong>The</strong> project team’s active participation in the open<br />
source community may be necessary for the team to use the<br />
technology. Based on an in-depth field study <strong>of</strong> industry s<strong>of</strong>tware<br />
engineering project students interacting with an open source<br />
community, we find that participation in the community may affect<br />
the team’s <strong>work</strong> <strong>and</strong> <strong>learning</strong> by strengthening the power <strong>of</strong> the<br />
broker between the team <strong>and</strong> the community. We outline pitfalls <strong>and</strong><br />
benefits <strong>of</strong> having student teams acquire development-related<br />
knowledge from open source communities. <strong>The</strong> findings are<br />
relevant to the organization <strong>and</strong> supervision <strong>of</strong> s<strong>of</strong>tware engineering<br />
student projects interacting with open source communities.<br />
Categories <strong>and</strong> Subject Descriptors<br />
K.3.2. [Computing Milieux]: <strong>Computer</strong> <strong>and</strong> Information Science<br />
Education<br />
General Terms: Human Factors<br />
Keywords: FLOSS, open source, s<strong>of</strong>tware engineering,<br />
s<strong>of</strong>tware engineering education, computer science education,<br />
communities <strong>of</strong> practice<br />
1. INTRODUCTION<br />
Student SE industry projects, also known as customer-driven<br />
projects, are designed to provide SE students with realistic<br />
experience from s<strong>of</strong>tware development <strong>work</strong>. <strong>The</strong> projects require<br />
that students relate to various stakeholders <strong>and</strong> engage in crosscommunity<br />
interaction [1]. Desired <strong>learning</strong> outcomes span both<br />
social <strong>and</strong> technical skills. <strong>The</strong> latter includes being flexible to<br />
customers’ technology requirements, which may be outside the<br />
team’s current experience <strong>and</strong> preference [2]. Through the<br />
pedagogical organization <strong>of</strong> SE project courses, student teams are<br />
provided with scaffolding for their <strong>learning</strong> process [3, 4], which<br />
might include guidance on the use <strong>of</strong> particular technology.<br />
Sometimes in the case <strong>of</strong> industry projects, neither course staff nor<br />
the customer is familiar with technology required in the project. <strong>The</strong><br />
Permission to make digital or hard copies <strong>of</strong> all or part <strong>of</strong> this <strong>work</strong> for<br />
personal or classroom use is granted without fee provided that copies are<br />
not made or distributed for pr<strong>of</strong>it or commercial advantage <strong>and</strong> that copies<br />
bear this notice <strong>and</strong> the full citation on the first page. To copy otherwise, or<br />
republish, to post on servers or to redistribute to lists, requires prior specific<br />
permission <strong>and</strong>/or a fee.<br />
ICSE’08, May 10-18, 2008, Leipzig, Germany.<br />
Copyright 2008 ACM 978-1-60558-079-1/08/05…$5.00.<br />
students have to demonstrate independence in coping with such<br />
situations.<br />
Requirements to learn <strong>and</strong> apply specific technologies may force SE<br />
teams to interact with user <strong>and</strong>/or developer communities through<br />
different degrees <strong>of</strong> community participation. For instance,<br />
involvement with open source s<strong>of</strong>tware (OSS) communities may be<br />
needed. In what follows, we will talk about open source s<strong>of</strong>tware<br />
communities for simplicity – tacitly assuming that similar<br />
considerations apply to free, libre <strong>and</strong> open source s<strong>of</strong>tware<br />
(FLOSS) communities.<br />
OSS communities are a relatively new arena for information<br />
acquisition <strong>and</strong> knowledge building in SE student projects.<br />
Supervisors might be in doubt about what is going on <strong>and</strong> what<br />
advice to give in such cases, feeling that their possibility to impact<br />
on the project process is diminished. On part <strong>of</strong> the students, having<br />
to participate in a different community in accordance with the<br />
practices <strong>of</strong> that community may result in increased knowledge as<br />
well as self confidence. Involvement with an OSS community in the<br />
pursuit <strong>of</strong> knowledge may thus benefit the SE student team.<br />
However, with an increased numbers <strong>of</strong> stakeholders to relate to, the<br />
complexity <strong>of</strong> project <strong>work</strong> [5] increases, <strong>and</strong> coordination becomes<br />
more challenging. Also, the introduction <strong>of</strong> unfamiliar, third party<br />
OSS components introduces difficulties in terms <strong>of</strong> providing a<br />
system design early in the SD life<strong>cycle</strong> [6]. This makes planning<br />
<strong>and</strong> estimation harder. Generally, there is very limited experience in<br />
student teams in respect <strong>of</strong> project management. Programmingrelated<br />
tasks, as will be argued, tend to be given priority, <strong>and</strong><br />
programmers given strong influence on the project process. We<br />
believe that project managers as well as supervisors <strong>of</strong> SE student<br />
teams may benefit from insights on how OSS participation can<br />
affect the projects, in order to be prepared to take adequate<br />
measures.<br />
With this as our point <strong>of</strong> departure, in this paper we explore why<br />
<strong>and</strong> how the interaction between a group <strong>of</strong> students, involved in an<br />
SE industry project, <strong>and</strong> an OSS developer community evolves over<br />
time. We further address how such interaction might influence the<br />
project process, particularly in respect <strong>of</strong> the power structures in the<br />
team.<br />
<strong>The</strong> paper is based on an in-depth exploratory study <strong>of</strong> the <strong>work</strong> <strong>and</strong><br />
<strong>learning</strong> <strong>of</strong> a student SE team during the spring term <strong>of</strong> 2007. We<br />
demonstrate how the interaction <strong>of</strong> the team with an OSS developer<br />
community can be paramount to the success <strong>of</strong> a SE project. In our<br />
case, we found that the OSS community readily responded to the<br />
students’ requests for assistance on the use <strong>of</strong> the OSS, <strong>and</strong> that<br />
issues <strong>of</strong> contribution from the team to the OSS community quickly<br />
arose. Also, we found that the team member responsible for<br />
791<br />
109
communicating with the OSS community through his brokering role<br />
maintained a position <strong>of</strong> power in the SE team.<br />
Based on our findings, we provide an account <strong>of</strong> (i) mechanisms<br />
enabling a SE team to successfully interact with an open source<br />
community <strong>and</strong> (ii) possible pitfalls <strong>and</strong> benefits <strong>of</strong> the engagement<br />
in terms <strong>of</strong> project success.<br />
First, we provide a theoretical background for the issues focused in<br />
our study. Next, we present our case, research method <strong>and</strong> findings.<br />
We further discuss the findings in the context <strong>of</strong> the above outlined<br />
research objectives, before concluding the paper.<br />
2. THEORY<br />
This section starts by providing a theoretical grounding for<br />
underst<strong>and</strong>ing power structures in SE teams, viewing teams as<br />
communities <strong>of</strong> practice. We then address how collaboration across<br />
communities can be achieved, before briefly outlining relevant<br />
research on how participation in OSS communities is typically<br />
structured <strong>and</strong> how OSS projects may serve as a resource in SE<br />
education.<br />
A student SE team can be viewed as a task-based <strong>learning</strong><br />
community [7], a form <strong>of</strong> community <strong>of</strong> practice [8] oriented<br />
towards <strong>learning</strong> <strong>and</strong> the time-limited effort <strong>of</strong> solving a particular<br />
task. <strong>The</strong> building <strong>of</strong> knowledge in the team is likely to be<br />
combined with the application, through trial-<strong>and</strong>-failure, <strong>of</strong> various<br />
solutions to the development task in question. Learning within the<br />
team may be based on expertise within the team, as seen in peer<br />
programming [9]. A student SE team may learn through interaction<br />
with course staff, e.g. a supervisor, <strong>and</strong> with the external customer,<br />
if the project is an industry project. In addition, there will typically<br />
be a need to acquire information from other sources, e.g. textbooks<br />
or resources on the Internet.<br />
Power structures in student SE teams can be seen as having several<br />
dimensions, for instance between the customer’s <strong>and</strong> the<br />
university’s goals, between experienced <strong>and</strong> inexperienced<br />
programmers, between various roles within the team, as well as<br />
between the team’s needs <strong>and</strong> individuals’ needs. A team typically<br />
regards it as the most important goal <strong>of</strong> their project to deliver<br />
<strong>work</strong>ing s<strong>of</strong>tware in line with the customer’s requirements. Even<br />
when the project report weighs heavily in the formal assessment <strong>and</strong><br />
grading <strong>of</strong> the project, documentation is typically conceived as<br />
something which is required by the university, but which is not too<br />
important for the customer – which might be a realistic observation<br />
[1] – <strong>and</strong> is not the real challenge <strong>of</strong> the project. With students’<br />
orientation towards the ‘real’ development task follows that<br />
programming is regarded as the core <strong>of</strong> the practice <strong>of</strong> the<br />
community constituted by the SE team. Accordingly, the skilled <strong>and</strong><br />
talented programmers get authority as core members <strong>of</strong> the team.<br />
<strong>The</strong> ongoing programming tasks tend to be both a major constraint<br />
<strong>and</strong> a driving force in the projects. As a result, whatever<br />
programmers decide to spend time on tends to be regarded as<br />
adequate use <strong>of</strong> time. Other team members, including the project<br />
manager, might not be able to judge whether it actually is. Team<br />
members good at activities regarded as secondary to programming<br />
will be closer to the periphery <strong>of</strong> the team community [10]. From<br />
the author’s pr<strong>of</strong>essional experience as SE project course staff over<br />
several years, it is common in the teams that the role <strong>of</strong> project<br />
manager is seen as important, but less central to the project than the<br />
programmer role, mainly because good programmers are a scarce<br />
resource.<br />
Sometimes, due to convenience or availability, information needed<br />
for the team’s <strong>work</strong> is collected from outside the communities <strong>of</strong> the<br />
team, the university, <strong>and</strong> the customer organization. <strong>The</strong><br />
information may be found in a more or less pedagogically organized<br />
form, <strong>and</strong> its gathering may be <strong>of</strong> a more or less interactive nature.<br />
Useful information sources include textbooks <strong>and</strong> tutorials,<br />
h<strong>and</strong>books, FAQs, case descriptions / examples, <strong>and</strong> templates.<br />
Often, such resources are freely available on the Internet <strong>and</strong> can be<br />
found by the aid <strong>of</strong> a search engine. More interactive sources <strong>of</strong><br />
information include communication channels belonging to a virtual<br />
community, e.g. associated with the use <strong>and</strong>/or development <strong>of</strong> a<br />
particular technology. Email lists, forums, wikis <strong>and</strong> blogs are<br />
frequently used to support virtual communities [11]. Also,<br />
synchronous communication may be supported through instant<br />
messaging <strong>and</strong> chat rooms. Gutwin et al. found that developers in<br />
OSS communities, who generally are distributed in their <strong>work</strong>,<br />
maintain awareness <strong>of</strong> each other primarily through text-based<br />
communication (mailing lists <strong>and</strong> chat systems) [12].<br />
In order to cross the border to a (virtual) community, a certain<br />
underst<strong>and</strong>ing <strong>of</strong> the rules <strong>of</strong> interaction <strong>and</strong> the community as a<br />
social system is required [13, 14]. Lave <strong>and</strong> Wenger described how<br />
the mechanism <strong>of</strong> legitimate peripheral participation (LPP) allows<br />
newcomers to gradually become members <strong>of</strong> a community <strong>of</strong><br />
practice [10]. Not only is LPP seen a way <strong>of</strong> entering the<br />
community; it is also seen as the basic way <strong>of</strong> <strong>learning</strong> the practices<br />
<strong>of</strong> the community. Wenger explained how two main mechanisms<br />
support the interaction between two communities: brokering <strong>and</strong><br />
boundary objects [8]. A broker is an individual who is a member <strong>of</strong><br />
both communities <strong>and</strong> is able to contribute to one community based<br />
on the underst<strong>and</strong>ing <strong>of</strong> the practices <strong>of</strong> the other community.<br />
Boundary objects, originally described by [15], are artifacts, abstract<br />
or concrete, that retain an identity across the communities they<br />
connect <strong>and</strong> play a role in the meaning-making <strong>of</strong> each community.<br />
A boundary object thus <strong>work</strong>s as a means for two communities to<br />
coordinate their practices, even if the communities underst<strong>and</strong> the<br />
boundary object in different ways <strong>and</strong> use it differently.<br />
Often, it will not be necessary for members <strong>of</strong> a SE team to become<br />
full-fledged members <strong>of</strong> a community to get necessary information<br />
from that community. In respect <strong>of</strong> communities related to a<br />
particular technology, there might be an opening for outsiders to<br />
participate as users, but not as developers.<br />
In the case <strong>of</strong> OSS development (OSSD), there is typically an<br />
openness that allows for participation in development for those who<br />
are able <strong>and</strong> willing to contribute. OSSD takes on many different<br />
shapes, but some general characteristics pertain to many developer<br />
communities [16]. <strong>The</strong> participation structure tends to be layered,<br />
with a core <strong>of</strong> active developers, intermediate layers <strong>of</strong> codevelopers,<br />
active users <strong>and</strong>, constituting the outer layer, passive<br />
users [17, 18]. Roughly, the model postulates that active users report<br />
defects, co-developers repair defects, <strong>and</strong> new <strong>and</strong> updated<br />
functionality is provided by core developers. Passive users may<br />
contribute by answering questions from other users. Contributions<br />
to the development <strong>of</strong> the OSS code may be seen as based a form <strong>of</strong><br />
gift economy where the giver achieves status from giving away. At<br />
the same time, these mechanisms assure a certain quality <strong>of</strong> the<br />
code [13]. <strong>The</strong> threshold for being admitted as a developer in an<br />
OSS community may vary between projects, large OSS projects<br />
based on the sponsorship <strong>of</strong> major corporations being far more<br />
restrictive than small OSS projects in terms <strong>of</strong> accepting<br />
development contributions.<br />
792<br />
110
Due to the openness <strong>of</strong> OSS communities, students may use them as<br />
an environment for <strong>learning</strong> which is interactive, ‘real’ <strong>and</strong> outside<br />
the university. OSS development has been acknowledged both as a<br />
means <strong>and</strong> as an objective <strong>of</strong> <strong>learning</strong> in SE education. Whereas the<br />
number <strong>of</strong> studies reporting lessons learned from students’<br />
participation in OSS development is still limited, experience <strong>and</strong><br />
lessons learned from student projects have been reported (e.g. [19,<br />
20]). Toth [20] lists the following benefits <strong>of</strong> using <strong>and</strong> extending<br />
OSS tools in student SE projects: a baseline <strong>of</strong> such tools<br />
automatically gives a ‘critical mass’ to the projects, the tools can<br />
freely be extended, using OSS is engaging for students, students<br />
may evolve their own tools (‘eating their own dog food’), <strong>and</strong><br />
students increase their value in the job market. Spinellis [21] points<br />
to the following benefits <strong>of</strong> <strong>work</strong>ing with OSS in general: increased<br />
industry contacts, pr<strong>of</strong>essional maturation, improved skills in<br />
written communication, experience with system administration, <strong>and</strong><br />
exposure to management approaches. <strong>The</strong>se are benefits equally<br />
relevant to SE <strong>and</strong> CS students as to SE pr<strong>of</strong>essionals. Jaccheri <strong>and</strong><br />
Østerlie [22] demonstrated that students may learn from an action<br />
research approach to participation in OSS development, resulting<br />
both in familiarity with action research <strong>and</strong> improved capabilities in<br />
programming <strong>and</strong> design. Ellis et al. [19] mention possible<br />
disadvantages <strong>of</strong> using OS projects as a basis for SE education: lack<br />
<strong>of</strong> documentation, inconsistent coding st<strong>and</strong>ards, <strong>and</strong> inconsistent<br />
quality.<br />
Lessons learned <strong>and</strong> guidelines emerging from current studies on<br />
student projects engaging in OSS communities generally address the<br />
pedagogical organization <strong>of</strong> SE project courses in which students’<br />
primary development task is itself OSS development. <strong>The</strong>re is<br />
however a lack <strong>of</strong> research on (i) ways in which the OSS<br />
community might be an important source <strong>of</strong> knowledge relevant to<br />
the students’ development task without being its main arena, <strong>and</strong> (ii)<br />
pedagogical issues arising from such a setting. This is where this<br />
paper aims to contribute.<br />
3. OUR CASE<br />
<strong>The</strong> particular SE project investigated in this <strong>work</strong> is part <strong>of</strong> an<br />
undergraduate level course at a Norwegian technical university. <strong>The</strong><br />
students take the course in the final semester <strong>of</strong> their Bachelor<br />
program, i.e. their 6 th semester. <strong>The</strong> students <strong>work</strong> in mostly selfformed<br />
teams <strong>of</strong> 3-5 students, developing s<strong>of</strong>tware for an external<br />
customer. Students make a prioritized wish list from a number <strong>of</strong><br />
available project descriptions mainly provided by industry, <strong>and</strong><br />
teams are assigned tasks by course staff according to students’<br />
preferences <strong>and</strong> skills. <strong>The</strong> <strong>work</strong>load <strong>of</strong> the course is half <strong>of</strong> the<br />
semester, but students tend to spend more time on the project, giving<br />
it higher priority than parallel courses. In addition to the s<strong>of</strong>tware<br />
product, the students must h<strong>and</strong> in a project report (both a<br />
preliminary <strong>and</strong> a final version) to the customer <strong>and</strong> course staff.<br />
Also, the teams give oral presentations <strong>of</strong> their projects halfway<br />
through the semester <strong>and</strong> at the end <strong>of</strong> the semester. <strong>The</strong>re are no<br />
lectures in the course except an introductory lecture in which<br />
available projects are presented. Each team receives supervision on<br />
their project process from a member <strong>of</strong> course staff, usually a<br />
teaching assistant. <strong>The</strong> customer is responsible for providing<br />
supervision on technical issues when necessary, as well as providing<br />
resources otherwise unavailable to the students (e.g. particular<br />
hardware, s<strong>of</strong>tware, <strong>and</strong> physical or virtual <strong>work</strong>space). Based on<br />
the course staff’s evaluation <strong>and</strong> customer feedback through an<br />
evaluation form, each team is given one grade (except in rare cases<br />
where major variation in individual <strong>work</strong> effort is documented by<br />
the team). <strong>The</strong> grade is based on a combination <strong>of</strong> the s<strong>of</strong>tware<br />
product, the documentation <strong>and</strong> the project process. <strong>The</strong> teams are<br />
provided with some templates <strong>and</strong> guidelines for their <strong>work</strong>, e.g.<br />
related to project planning, status reporting <strong>and</strong> contents <strong>of</strong> the<br />
project report.<br />
<strong>The</strong>re are no general requirements that any particular process<br />
model, analysis/design methodology, development tool or<br />
collaboration technology be used in the projects. Students are<br />
expected to relate to the customer’s requirements <strong>and</strong> draw on their<br />
skills <strong>and</strong> resources from previous <strong>and</strong> parallel courses at the<br />
university. <strong>The</strong> teams use tools available at the university, provided<br />
by the customer, or otherwise available. When needed <strong>and</strong> feasible<br />
within the time frame <strong>of</strong> the project, the teams are expected to get<br />
into new technology as part <strong>of</strong> their project <strong>work</strong>.<br />
<strong>The</strong> student team followed in the case study presented in this paper<br />
consisted <strong>of</strong> five male students: Ethan, George, Sam, Owen, <strong>and</strong><br />
Morgan (names have been altered), all in their twenties. <strong>The</strong>ir<br />
development task was to make an auctioning system for an IT<br />
consultancy company, here called Anniva. <strong>The</strong> company wanted to<br />
use the system for in-house purposes <strong>and</strong> as a substitute for an<br />
existing system used by employees to sell items (such as PCs <strong>and</strong><br />
bi<strong>cycle</strong>s) to colleagues on a private basis. As a secondary objective<br />
for the project, the company wanted the team to make use <strong>of</strong> a<br />
particular open source Java development frame<strong>work</strong>, here denoted<br />
PLENTI, in making the auctioning system, to see how the<br />
frame<strong>work</strong> could be utilized. <strong>The</strong> team succeeded in developing the<br />
auctioning system, receiving the grade B on their project. (B is<br />
regarded as a good grade; <strong>of</strong> the nine project teams in the 2007<br />
semester <strong>of</strong> the course, there were 2 As, 4 Bs <strong>and</strong> 3 Cs.) Crucial to<br />
the team’s development <strong>work</strong>, as will be seen, was their interaction<br />
with the PLENTI developer community.<br />
4. RESEARCH METHOD<br />
<strong>The</strong> student team <strong>of</strong> our study was observed during the spring<br />
semester 2007, i.e. over a period <strong>of</strong> four months. Given the real-life<br />
context, focus on a contemporary phenomenon, lack <strong>of</strong> control over<br />
events <strong>and</strong> our intent to explore “how” <strong>and</strong> “why” aspects <strong>of</strong> the<br />
setting, a case study was appropriate [23]. Part-time participant<br />
observation was conducted with adherence to principles <strong>of</strong><br />
interpretive field research [24]. For the research project at large, data<br />
(mainly documents <strong>and</strong> interviews) was collected across all teams in<br />
the course. Data for the study reported in this paper originate from<br />
an in-depth study <strong>of</strong> one particular project team <strong>and</strong> includes<br />
recordings, field notes <strong>and</strong> pictures from 15 hours <strong>of</strong> meetings<br />
between the team <strong>and</strong> customer/supervisor <strong>and</strong> 60 hours <strong>of</strong> teaminternal<br />
<strong>work</strong> sessions <strong>and</strong> meetings (mainly in the computer lab),<br />
copies <strong>of</strong> pages from the team’s wiki, some logged msn<br />
conversations, various project documents from the team’s<br />
<strong>work</strong>space, the preliminary <strong>and</strong> final version <strong>of</strong> the project report,<br />
all email correspondence going through the team’s email list,<br />
recorded interviews with the team, the supervisor <strong>and</strong> the customer,<br />
<strong>and</strong> the threads in the PLENTI user forum (listserv) resulting from<br />
the team’s requests to the forum.<br />
Generally, the researcher was given free access to <strong>work</strong> sessions <strong>and</strong><br />
meetings (participation <strong>and</strong> recording) as well as to the team’s<br />
server <strong>work</strong>space, wiki <strong>and</strong> email list. <strong>The</strong> researcher attended all<br />
meetings with the customer, with the exception <strong>of</strong> a couple <strong>of</strong><br />
meetings which were audio recorded for the researcher by a team<br />
793<br />
111
member. Field notes were made during <strong>and</strong> immediately after the<br />
sessions, to the extent possible. <strong>The</strong> team’s supervisor was<br />
interviewed during <strong>and</strong> after the project, <strong>and</strong> the team’s customer<br />
was interviewed after the project. <strong>The</strong> team was interviewed once<br />
during the middle <strong>of</strong> the project <strong>and</strong> once after completion <strong>of</strong> the<br />
project.<br />
Most <strong>of</strong> the meeting recordings <strong>and</strong> interviews have been fully<br />
transcribed. Recordings from <strong>work</strong> sessions have been partly<br />
examined <strong>and</strong> summarized based on their perceived relevance in the<br />
analysis process. Quotations have been translated into English by<br />
the researcher at need. Analysis <strong>of</strong> the data started out with some<br />
theoretical concepts being perceived as central (cf. the Principle <strong>of</strong><br />
Abstraction <strong>and</strong> Generalization, [24]). Coding was performed on the<br />
basis <strong>of</strong> these concepts, by the aid <strong>of</strong> a computerized qualitative<br />
analysis tool (Atlas.ti). New codes were added as themes developed<br />
during interpretation <strong>and</strong> writing. For instance, PLENTI was used as<br />
a code to mark <strong>and</strong> organize data related to the team’s use <strong>of</strong> the<br />
frame<strong>work</strong> <strong>and</strong> interaction with the community. Further, the data<br />
have been chronologically structured within selected themes.<br />
<strong>The</strong> researcher was the coordinator <strong>of</strong> the project course, grading<br />
the projects, but not supervising any group. In the process <strong>of</strong><br />
observing the Anniva team, great care was taken to avoid mixing up<br />
the roles <strong>of</strong> researcher <strong>and</strong> course staff <strong>and</strong> to be aware <strong>of</strong> the<br />
possible effects <strong>of</strong> ‘the teacher being present’ (cf. the Principle <strong>of</strong><br />
Interaction Between the Researcher <strong>and</strong> the Subjects, [24]). As part<br />
<strong>of</strong> the initial agreement between the researcher <strong>and</strong> the team, it was<br />
decided that another member <strong>of</strong> course staff set the grade on their<br />
project. Further, it was agreed that the researcher was not to<br />
supervise them, but restrain interaction to being social without<br />
causing too much disturbance. <strong>The</strong> reason for not supervising was<br />
not mainly that supervision would influence the case. <strong>The</strong> influence<br />
<strong>of</strong> the researcher is an inevitable consequence <strong>of</strong> participant<br />
observation, even if the degree <strong>of</strong> influence can, <strong>and</strong> should, be<br />
limited. <strong>The</strong> team was not, however, to be given advantages as<br />
compared to the other teams in the course, or at least the sum <strong>of</strong><br />
advantages <strong>and</strong> disadvantages <strong>of</strong> being research subjects should be<br />
perceived as close to zero by the team <strong>and</strong> by the other teams. <strong>The</strong>se<br />
considerations impacted on the possibility to have the researcher’s<br />
interpretation validated with the research subjects during the study<br />
(cf. the Principle <strong>of</strong> Multiple Interpretations, [24]). As all teams in<br />
the course were however interviewed about their project <strong>work</strong> <strong>and</strong><br />
viewpoints on the course in the middle <strong>of</strong> the semester, <strong>and</strong> on that<br />
occasion it was possible to have some prompted viewpoints from<br />
the Anniva team. Also, the day after the completion <strong>of</strong> the project, a<br />
three hour interview was conducted with the team to have their<br />
feedback on the researcher’s preliminary interpretation <strong>of</strong> the<br />
project.<br />
In reflective discussions within the team towards the end <strong>of</strong> the<br />
project <strong>and</strong> in the interview conducted after the project, the team<br />
expressed that they had perceived the researcher as non-interfering<br />
<strong>and</strong> agreeable, <strong>and</strong> that they had not received project supervision<br />
from her.<br />
To validate findings reported in this paper, a draft version has been<br />
reviewed by two <strong>of</strong> the team members, Owen <strong>and</strong> Ethan. Ethan<br />
expressed that he liked the paper. Owen provided more detailed<br />
viewpoints which have been incorporated in the Analysis <strong>and</strong><br />
Findings <strong>and</strong> Discussion sections.<br />
5. ANALYSIS AND FINDINGS<br />
In presenting the findings <strong>of</strong> our study in terms <strong>of</strong> the team’s<br />
interaction with the PLENTI community, we have made a<br />
chronological structuring <strong>of</strong> the material, distinguishing what we see<br />
as four partly overlapping phases.<br />
<strong>The</strong> first phase is a (dis)orientation phase in which the team strives<br />
to underst<strong>and</strong> what PLENTI is <strong>and</strong> how to obtain adequate<br />
knowledge about the frame<strong>work</strong>. In the second phase, PLENTI is<br />
regarded by the team as an obstacle, but starts taking on the role <strong>of</strong> a<br />
tool (albeit not quite mastered by the team). In the third phase, the<br />
team is communicating with PLENTI users <strong>and</strong> developers through<br />
the listserv <strong>of</strong> the community, actively making use <strong>of</strong> the frame<strong>work</strong><br />
as a tool in the project development task. In the fourth phase signs<br />
<strong>of</strong> a feeling <strong>of</strong> identity related to the PLENTI community<br />
participation can be found in the team, <strong>and</strong> the team is contributing<br />
to the PLENTI development.<br />
In the subsequent discussion (Section 6), we will argue that the<br />
transition to the third phase was essential to the project, <strong>and</strong> address<br />
how this transition came about in our case. In the present section,<br />
we characterize each phase in more detail, giving some illustrative<br />
excerpts from the data where appropriate.<br />
5.1.1 First phase (January-February): What is<br />
PLENTI?<br />
In the original project description as well as in the initial customer<br />
meeting, it was specified by the customer that PLENTI should be<br />
used as a development frame<strong>work</strong>. <strong>The</strong> reason was that one <strong>of</strong> the<br />
two customer representatives had attended a seminar with an<br />
enthusiastic PLENTI developer <strong>and</strong> thus caught interest in the<br />
frame<strong>work</strong>. During the first customer meeting, the team was <strong>of</strong>fered<br />
the alternative <strong>of</strong> using another frame<strong>work</strong>, but not PHP, which was<br />
the only alternative with which they were familiar. This left<br />
PLENTI as the de facto option. Adding to the arguments to use<br />
PLENTI, the customer expressed interest in having it tried out to see<br />
how it <strong>work</strong>ed for this type <strong>of</strong> application development.<br />
None <strong>of</strong> the students in the team had any advance knowledge about<br />
PLENTI. Further, no one had any experience with the programming<br />
<strong>of</strong> web applications with servlets. <strong>The</strong> web pages <strong>of</strong> the PLENTI<br />
community were the only resource about the frame<strong>work</strong> known to<br />
the team.<br />
<strong>The</strong> team decided that every team member was to learn PLENTI<br />
<strong>and</strong> participate in the programming tasks. Two team members, Sam<br />
<strong>and</strong> Owen, were assigned the task <strong>of</strong> starting the knowledge<br />
acquisition about PLENTI. Ethan <strong>and</strong> George, accepted as the lead<br />
programmers in the team, took on the task <strong>of</strong> developing a prototype<br />
to demonstrate to the customer as a basis for refining the functional<br />
requirements. Ethan <strong>work</strong>ed on the web user interface whereas<br />
George focused on the underlying business logic, none <strong>of</strong> them<br />
making use <strong>of</strong> PLENTI for their tasks. Both programmers<br />
participated in team-internal discussions on the role <strong>and</strong> use <strong>of</strong><br />
PLENTI, though. <strong>The</strong> fifth team member, Morgan, took the role as<br />
project manager, which made him mainly responsible for overall<br />
project planning, generating <strong>and</strong> h<strong>and</strong>ing in bi-weekly status reports<br />
<strong>and</strong> activity plans to the supervisor, scheduling meetings, <strong>and</strong><br />
making sure that <strong>work</strong> on the project report progressed in<br />
accordance with the deadlines.<br />
794<br />
112
5.1.2 Second phase (February-May): PLENTI as an<br />
obstacle gradually becoming a tool<br />
It soon became evident to the team that getting into the <strong>work</strong>ings <strong>of</strong><br />
the frame<strong>work</strong> was not an easy task. Sam <strong>and</strong> Owen spent many<br />
hours in the PC lab reading through descriptions <strong>and</strong> examples, Sam<br />
doing some exploratory coding <strong>and</strong> Owen reflecting overview level<br />
information on the frame<strong>work</strong> in the S<strong>of</strong>tware Development Plan<br />
which had to be h<strong>and</strong>ed in early in February.<br />
<strong>The</strong> web documentation <strong>of</strong> the frame<strong>work</strong> turned out to rely on the<br />
reader’s prior knowledge <strong>of</strong> the development <strong>of</strong> web applications,<br />
which the team lacked. <strong>The</strong> examples provided in the material were<br />
not relevant enough for the team to be able to easily draw on them.<br />
<strong>The</strong> team explained their challenges with PLENTI to the supervisor<br />
<strong>and</strong> the customer in meetings with them.<br />
Exhibit 1: Excerpt from February 16, meeting between the team<br />
<strong>and</strong> the customer:<br />
Ethan: ”We had a bit <strong>of</strong> trouble with, with.. that is, we have a small<br />
problem with the PLENTI documentation. It is a bit.. <strong>The</strong>y have a<br />
lot <strong>of</strong> examples, but there is sort <strong>of</strong> nothing that is.. the most basic<br />
stuff. <strong>The</strong>y assume that you know quite a lot.” Owen:”<strong>The</strong> examples<br />
<strong>and</strong> documentation found there are mainly presentations, mostly <strong>of</strong><br />
how PLENTI is better than earlier <strong>and</strong> similar tools. We have some<br />
problems with, sort <strong>of</strong>.. the totally basic stuff, they aren’t explained<br />
anywhere, so it becomes a lot <strong>of</strong> trying <strong>and</strong> failing.”<br />
Ethan <strong>and</strong> George continued focusing on their preferred parts <strong>of</strong> the<br />
prototyping <strong>work</strong>, <strong>and</strong> in the larger part <strong>of</strong> this phase, the team had<br />
little progress with their underst<strong>and</strong>ing <strong>of</strong> PLENTI.<br />
Gradually some knowledge <strong>of</strong> the <strong>work</strong>ings <strong>of</strong> PLENTI developed,<br />
however, as the programmers started using the frame<strong>work</strong> in their<br />
prototyping. <strong>The</strong> team was however aware that they did not yet<br />
utilize the powerful mechanisms that would be the main reason for<br />
using the frame<strong>work</strong>. Attempts to do the latter were postponed,<br />
awaiting the completion <strong>and</strong> demonstrating <strong>of</strong> the prototypes. Also,<br />
<strong>work</strong> on design <strong>and</strong> architecture models <strong>of</strong> the auctioning system<br />
was postponed, again with reference to the lack <strong>of</strong> underst<strong>and</strong>ing <strong>of</strong><br />
how PLENTI should be used. In the eyes <strong>of</strong> Ethan, it would be more<br />
or less meaningless to draw a design model: “It would just be a big<br />
box named PLENTI with a number <strong>of</strong> arrows coming out <strong>of</strong> it”.<br />
In the midterm version <strong>of</strong> the project report, h<strong>and</strong>ed in on March 7,<br />
as well as in the related oral presentation, PLENTI was still<br />
presented by the team as something they needed to learn properly.<br />
<strong>The</strong> auctioning system prototype demonstrated was mostly a mockup<br />
<strong>and</strong> hardly made use <strong>of</strong> any <strong>of</strong> the functionality in the<br />
frame<strong>work</strong>.<br />
In the customer meeting on March 28, it was clear to the customer<br />
as well as to the team that the progress <strong>of</strong> the project was not<br />
satisfactory.<br />
Exhibit 2: Excerpt from March 28, meeting between the team<br />
<strong>and</strong> the customer:<br />
Ethan: ”Ok, we have made several prototypes, so that.. we have<br />
learnt PLENTI by programming. But we haven’t made any<br />
completely robust solutions yet.” Owen: ”And then we have used a<br />
lot <strong>of</strong> time to learn, study PLENTI, then, because the documentation<br />
is quite bad, I think.”<br />
5.1.3 Third phase (April-May) Gradual PLENTI<br />
community participation<br />
In the customer meeting on March 28. Ethan reports “having been a<br />
bit on the IRC” communicating with the PLENTI community,<br />
reporting that there was “this guy” that he communicated with.<br />
Exhibit 3: Excerpt from March 28, meeting between the team<br />
<strong>and</strong> the customer. Ethan talks, quoting his IRC communication<br />
with Bernhard:<br />
’Why do you do it that way?’ And we just: ‘What? That’s how we do it in JSP.’ ’Don’t do it that way!’ <br />
No good, no good, so…”<br />
Expressing the above, explaining how Bernhard had reacted<br />
negatively to their approach, Ethan still appeared satisfied with the<br />
IRQ dialogue. He had in fact reached a PLENTI developer with his<br />
request, <strong>and</strong> the dialogue resulted in relevant directions for the team.<br />
After the Easter holiday, following up on the success with IRC,<br />
Ethan turned to the PLENTI users forum, a listserv service linked to<br />
the PLENTI web pages. <strong>The</strong> subsequent interaction between Ethan<br />
<strong>and</strong> the community in this forum amounted to eight threads, most <strong>of</strong><br />
which with several postings, during the course <strong>of</strong> the project.<br />
A condensed thread from the forum is shown in Exhibit 4. Other<br />
requests from Ethan were addressed by Bernhard partly in parallel,<br />
in different threads.<br />
Exhibit 4: Excerpt from April 13.-18., PLENTI users forum on<br />
the Internet:<br />
April 13.: Ethan: ”We’re trying to make a credentials manager for<br />
authenticating against LDAP. We’ve written some code but we’re<br />
unsure about what to do next, or if the code is correct. Any ideas? ..<br />
Code: ”<br />
April 14.: BG “Hi Ethan, using the upcoming PLENTI 1.6, it’s very<br />
easy to plug in your own credentials manager ”<br />
April 16. Ethan: “Hi, Bernhard, In addition to , what files do<br />
we need? ”<br />
April 18.: BG: “Just should suffice ”<br />
April 18.: Ethan: “Thank you Bernhard! That <strong>work</strong>ed perfectly. We<br />
have now functioning LDAP authentication in our webapp. Has<br />
integrating LDAP authentication directly in PLENTI been<br />
considered? Sincerely, Ethan Lake”<br />
April 18.: BG: “Wonderful! Several people have asked for it in the<br />
past, it would be a very welcome contribution to the project.”<br />
Whereas the communication shown in Exhibit 4 took place in the<br />
PLENTI users forum (as opposed to the developer forum, which<br />
also could be found on the web site), changes to PLENTI itself were<br />
discussed in the last two postings. <strong>The</strong> initial posting was the first<br />
request from Ethan on the forum. Bernhard, the chief developer <strong>of</strong><br />
PLENTI, was, with one exception, the one who answered Ethan’s<br />
(<strong>and</strong> a number <strong>of</strong> other users’) requests on the listserv. He gave<br />
quick response, <strong>of</strong>ten on the same day (or night). <strong>The</strong>se observations<br />
together may be taken as an indication that the OSS community is a<br />
very small one, with small distance between user <strong>and</strong> developer as<br />
well as between user issues <strong>and</strong> development issues.<br />
At one point, Ethan was asked by Bernhard to send large pieces <strong>of</strong><br />
code by private email instead <strong>of</strong> pasting the code directly into the<br />
forum. Ethan sent code by email as requested. On a couple <strong>of</strong><br />
795<br />
113
occasions, Bernhard asked for more information or clarification to<br />
underst<strong>and</strong> Ethan’s questions, on which Ethan immediately<br />
followed up. Generally, communication between the two appeared<br />
effective, to-the-point, polite <strong>and</strong> friendly.<br />
<strong>The</strong> interaction with the PLENTI community was conveyed to the<br />
team through the team’s mailing list. Ethan made sure that his<br />
requests <strong>and</strong> Bernhard’s answers were in this way shared with the<br />
rest <strong>of</strong> the team. <strong>The</strong> new knowledge was used by the programmers<br />
- Ethan, George <strong>and</strong> Sam - to develop <strong>and</strong> modify their code. In<br />
discussions within the student team, new issues were identified that<br />
required more consultation with Bernhard. Ethan took on the job <strong>of</strong><br />
formulating <strong>and</strong> submitting postings, which he sometimes did in the<br />
PC lab <strong>and</strong> sometimes in his home at night. Ethan’s role in this was<br />
never questioned by the team.<br />
In terms <strong>of</strong> the division <strong>of</strong> project <strong>work</strong>, Owen produced use cases<br />
<strong>and</strong> other parts <strong>of</strong> the documentation, coordinating with the others<br />
as needed. Morgan continued performing project management,<br />
more by following up on what the others were in fact doing than by<br />
pushing the process. Sam largely assisted with programming tasks.<br />
Ethan <strong>and</strong> George were in effect driving the process, everyone<br />
realizing that their success with the coding was essential to the team.<br />
Morgan <strong>and</strong> Owen took on the main responsibility <strong>of</strong> making the<br />
final report, leaving the design <strong>and</strong> programming-related contents to<br />
the programmers. Until the last weeks <strong>of</strong> the project, <strong>and</strong> with the<br />
exception <strong>of</strong> a data model which was created early in the project <strong>and</strong><br />
had been heavily modified after the customer meeting on March 28.<br />
the creation <strong>of</strong> models <strong>and</strong> diagrams to be used in the project report<br />
was largely ignored by Ethan <strong>and</strong> George, who were mostly<br />
concerned about coding <strong>and</strong> debugging. Ethan <strong>and</strong> George were<br />
regarded as the ones having control <strong>of</strong> the totality <strong>of</strong> the system<br />
under development. George was highly estimated for his<br />
programming skills. <strong>The</strong> rest <strong>of</strong> the team, in his absence, referred to<br />
how they needed him to get going on particular tasks for the project<br />
to succeed.<br />
Ethan made the architecture model asked for on several occasions<br />
by customer <strong>and</strong> supervisor. Suggestions to Ethan from Owen <strong>and</strong><br />
Sam in the direction <strong>of</strong> a layered design model were turned down by<br />
Ethan, who continued to regard PLENTI as somehow noncompliant<br />
with a layered architecture <strong>and</strong> came up with a model<br />
following his own notation.<br />
<strong>The</strong> following comment to a draft version <strong>of</strong> this paper was made by<br />
Owen: “In the absence <strong>of</strong> a thorough elaboration phase in which one<br />
builds a well defined s<strong>of</strong>tware architecture <strong>and</strong> decide on the design<br />
<strong>of</strong> the system, a lot <strong>of</strong> the development <strong>and</strong> design <strong>of</strong> the product<br />
was up to those actually writing the larger part <strong>of</strong> the code”.<br />
5.1.4 Fourth phase (April-May): Contribution to<br />
PLENTI<br />
<strong>The</strong> point at which Ethan started discussing with Bernhard about<br />
possible contributions to PLENTI might be seen as the start <strong>of</strong> the<br />
phase in which the team were seen as, or saw themselves as,<br />
contributors to the PLENTI community.<br />
Three main events constituted the contribution part <strong>of</strong> the team’s<br />
interaction with PLENTI. First, there was a nightly build <strong>of</strong><br />
PLENTI in which the changes directly resulted from requests from<br />
the Anniva team <strong>and</strong> a bug subsequently discovered by Bernhard.<br />
Second, the team wanted to integrate the system with a MySQL<br />
database, for which there was no support in the frame<strong>work</strong>. In this<br />
case, Bernhard indicated how long it would take to implement the<br />
functionality (i.e., if they wanted to do it); at least a week’s <strong>work</strong>.<br />
<strong>The</strong> team took this to mean a lot more <strong>work</strong> for them as<br />
inexperienced programmers, <strong>and</strong> the idea was ab<strong>and</strong>oned. Third,<br />
there was the LDAP functionality referred to in Exhibit 4, suggested<br />
by Ethan <strong>and</strong> applauded by Bernhard. Even after the completion <strong>of</strong><br />
the project, the team continued discussing the possibility <strong>of</strong><br />
contributing such a module. That development job was however<br />
never carried out.<br />
<strong>The</strong> team’s account as given in the common <strong>reflection</strong> note shows<br />
how PLENTI is given a very central role in the team’s<br />
underst<strong>and</strong>ing <strong>of</strong> their project process. We note that the frame<strong>work</strong><br />
still plays a role as an excuse <strong>and</strong> explanation for insufficient project<br />
management.<br />
Exhibit 5: Excerpt from the common <strong>reflection</strong> note in the final<br />
project report<br />
What did not <strong>work</strong> so well in the project when it comes to project<br />
management? It has been problematic to plan the use <strong>of</strong> time in a<br />
realistic way. Our lack <strong>of</strong> experience with PLENTI <strong>and</strong> web<br />
application development resulted in frequent ad-hoc estimates. This<br />
was particularly difficult in the beginning <strong>of</strong> the project. During the<br />
development process we naturally gained more insight into PLENTI<br />
<strong>and</strong> web application development, which made it easier to make<br />
realistic estimates.<br />
In respect <strong>of</strong> achieving good project management: What would<br />
you do differently if you later on participate in a similar project?<br />
If we had <strong>work</strong>ed more intensively in the beginning <strong>of</strong> the project<br />
with our underst<strong>and</strong>ing <strong>of</strong> PLENTI, <strong>and</strong> the more advanced<br />
functions that we ended up using, project management <strong>and</strong><br />
scheduling would probably have been a bit easier.<br />
<strong>The</strong> formerly mentioned period <strong>of</strong> inactivity was redeemed by a<br />
good customer meeting, <strong>and</strong> this might have been held earlier. Also<br />
we could have used the PLENTI email list earlier, as we got good<br />
user support from the PLENTI developers there.<br />
In the final oral presentation on May 25, the team spent much time<br />
describing the role <strong>of</strong> PLENTI in their project: how PLENTI had<br />
caused trouble, how the solution was based on PLENTI, <strong>and</strong> how<br />
interaction with the PLENTI developers had been essential. A<br />
certain pride with the process was observable in the team.<br />
In the individual <strong>reflection</strong> note, written by each team member as a<br />
supplement to the common <strong>reflection</strong> note, Ethan does not make a<br />
point <strong>of</strong> his role in respect <strong>of</strong> PLENTI when explaining his role <strong>and</strong><br />
tasks in the project. He mentions his main responsibility for coding<br />
related to the user interface (CSS <strong>and</strong> HTML), for making diagrams,<br />
<strong>and</strong> for reviewing <strong>and</strong> addressing language issues in the report. In<br />
the note, responding to the question about whether he would have<br />
liked to have a different role in a similar project on a later occasion,<br />
he states that: “I am very happy about what I have been <strong>work</strong>ing<br />
with in the project”.<br />
6. DISCUSSION<br />
From the chronological presentation <strong>of</strong> findings, we note that the<br />
transition from the second to the third phase <strong>of</strong> interaction with the<br />
PLENTI community was essential to the project. It was in the third<br />
phase they got access to the knowledge they needed to be able to<br />
utilize the frame<strong>work</strong>. <strong>The</strong> fourth phase, contributing to the<br />
development <strong>of</strong> PLENTI, was not strictly necessary for the project<br />
team to succeed with their development task.<br />
796<br />
114
Getting from the second to the third phase can be seen mostly as a<br />
result <strong>of</strong> the team’s realization that progress was too slow <strong>and</strong> that<br />
something must be done. <strong>The</strong> means for the transition to take place,<br />
was the interaction with the OSS community, the only place where<br />
the necessary knowledge could be found. In what follows, we will<br />
focus on an important mechanism enabling this interaction: the<br />
brokering performed by one team member.<br />
In our case, Ethan took on the role <strong>of</strong> broker between the team <strong>and</strong><br />
the OSS community. In the community, he presented the team’s<br />
challenges in a way which was meaningful to the OSS developers.<br />
Back in the SE student team, Ethan explained what Bernhard’s<br />
feedback meant to their <strong>work</strong> <strong>and</strong> translated it into Java code in<br />
collaboration with the other programmers.<br />
In the brokering process, there were artifacts that can be seen as<br />
boundary objects [15] enabling meaning-making across<br />
communities. <strong>The</strong> auctioning system, from the outset unknown to<br />
Bernhard, gradually became more useful in mediating interaction as<br />
specific modules were discussed <strong>and</strong> changed. Pieces <strong>of</strong> code were<br />
described but also submitted in ‘raw’ format, explored, <strong>and</strong> perhaps<br />
modified <strong>and</strong> returned. Further, the PLENTI frame<strong>work</strong> itself<br />
mediated the interaction: From the start, PLENTI was well known<br />
to Bernhard but not sufficiently mastered by Ethan <strong>and</strong> the team. As<br />
the interaction proceeded, the team’s knowledge improved to the<br />
point that it became meaningful to discuss changes to PLENTI<br />
itself. <strong>The</strong> roles <strong>of</strong> the boundary objects in each <strong>of</strong> the communities<br />
were <strong>of</strong> course different: For the SE team, PLENTI was a means to<br />
develop a <strong>work</strong>ing auctioning system. To Bernhard, having<br />
someone doing development by the use <strong>of</strong> PLENTI was a means to<br />
improve PLENTI itself.<br />
We leave the elaboration on boundary objects at this point, noting<br />
that they should be seen as integral to the brokering <strong>and</strong> enriching<br />
our underst<strong>and</strong>ing <strong>of</strong> the process. Having argued that Ethan was in<br />
fact a broker, we turn the discussion to how the brokering empowers<br />
the broker, with possible effects on the SE project.<br />
6.1 <strong>The</strong> power <strong>of</strong> the broker<br />
We suggest that the broker between a SE team <strong>and</strong> an OSS<br />
community gains increased power in the team in three ways: 1)<br />
through his position as gatekeeper <strong>of</strong> knowledge needed by the<br />
team, 2) through strengthening his existing authority as a<br />
programmer, <strong>and</strong> 3) through increased credibility in both<br />
communities through the feedback provided in the OSS community<br />
In what follows, we pursue these arguments referring to our case.<br />
6.1.1 <strong>The</strong> power <strong>of</strong> the gatekeeper<br />
As the broker translates knowledge from one community to the<br />
other, he simultaneously controls which information flows between<br />
the communities. It is the broker’s interpretation which will reach<br />
each community, even if he faithfully aims to represent the<br />
community’s interests <strong>and</strong> not his own vested ones. When the<br />
receiving community depends on the information to succeed with<br />
their core activities, controlling the information is a great source <strong>of</strong><br />
power <strong>and</strong> an opportunity to control the agenda. Some<br />
characteristics <strong>of</strong> our case point to the gatekeeping role <strong>of</strong> the<br />
broker.<br />
First, it is only Ethan in the team that interacts with the PLENTI<br />
community. Thus, no one else actually passes the only gate towards<br />
the PLENTI community, namely the listserv, apart from reading the<br />
resulting threads there <strong>and</strong> taking part in the interpretation <strong>of</strong> the<br />
new knowledge in the context <strong>of</strong> the team’s <strong>work</strong>. Nothing would<br />
have prevented the other team members from communicating with<br />
Bernhard if they wanted to. However, the division <strong>of</strong> labour letting<br />
Ethan take on the role as broker <strong>and</strong> gatekeeper seems to be<br />
generally agreed-upon in the team, perhaps because they<br />
acknowledged his abilities to attend to the interests <strong>of</strong> the team.<br />
Second, Ethan is fairly active in taking <strong>and</strong> keeping the role as<br />
broker, thus in a sense defending it as his domain. For instance, in<br />
meetings, when unresolved, PLENTI-related issues arise, he is<br />
quick at suggesting that he pose a request to the community.<br />
Third, Ethan’s communication with the PLENTI community is done<br />
through postings partly sent outside the hours <strong>of</strong> collocated <strong>work</strong> in<br />
the SE team. <strong>The</strong> questions asked <strong>and</strong> the interpretations made in<br />
the OSS community are therefore withheld from, or at least<br />
distanced from the collaborative activity <strong>of</strong> the team. No one in the<br />
team ever expressed any dissatisfaction over this. This might point<br />
to the team’s general appreciation <strong>of</strong> members’ flexibility <strong>of</strong> <strong>work</strong><br />
(time <strong>and</strong> place), but may also indicate satisfaction with the way<br />
Ethan reflected their discussions <strong>and</strong> represented their interests in<br />
the OSS community.<br />
6.1.2 <strong>The</strong> programmer increasing his power<br />
<strong>The</strong> role as broker towards a developer community must be held by<br />
an individual having enough knowledge about development to be<br />
able to effectively communicate over issues important to the<br />
community. In practice, this means a programmer. As argued in<br />
section 2, programming skills give authority in the SE teams. <strong>The</strong><br />
programmer brokering towards a developer community gets more<br />
programming skills <strong>and</strong> thus more power.<br />
<strong>The</strong> following points from our case serve to illustrate how being a<br />
programmer is highly central to the broker:<br />
First, in the OSS community, the language <strong>of</strong> interaction is largely<br />
source code as well as considerations on aspects <strong>of</strong> source code, on<br />
the level <strong>of</strong> single lines <strong>of</strong> code <strong>and</strong> on the level <strong>of</strong> modules <strong>and</strong><br />
versions <strong>of</strong> the entire system. Ethan makes his requests<br />
underst<strong>and</strong>able to Bernhard by giving excerpts from their source<br />
code, <strong>and</strong> receives answers directly addressing aspects <strong>of</strong> the code<br />
or even containing modifications <strong>of</strong> the code itself. In addition to<br />
being able to ‘speak code’, there is a need to follow the implicit<br />
rules <strong>of</strong> communication in the forum. It is an open question whether<br />
programmers generally have a better starting point for<br />
communicating in virtual developer forums, but it is not unlikely<br />
that many programmers are experienced in gathering information<br />
from this type <strong>of</strong> site. In all his contributions to the PLENTI users’<br />
forum, Ethan demonstrates a high degree <strong>of</strong> literacy in terms <strong>of</strong> this<br />
communication channel. His contributions appear friendly, polite,<br />
concise <strong>and</strong> goal-directed, <strong>and</strong> always appropriate to the context.<br />
Second, being a broker, you need to represent your community.<br />
Representing a SE team in the context <strong>of</strong> a developer community<br />
means being accountable for the s<strong>of</strong>tware produced by the team. It<br />
is difficult to see how team members other than Ethan <strong>and</strong> George,<br />
who had the h<strong>and</strong>s-on experience with the current version <strong>of</strong> the<br />
code <strong>and</strong> its unresolved bugs as well as the overview <strong>of</strong> the current<br />
version <strong>of</strong> the system, could have communicated with PLENTI<br />
developers over development issues. In face-to-face meetings with<br />
project stakeholders, e.g. the customer, Ethan frequently took on the<br />
task <strong>of</strong> articulating the status <strong>of</strong> <strong>work</strong> <strong>and</strong> the current issues <strong>and</strong><br />
priorities <strong>of</strong> the team. This <strong>work</strong>ed only partially, as Ethan’s role<br />
<strong>and</strong> conduct was not quite accepted by the customer. In the OSS<br />
797<br />
115
community, however, Ethan always convincingly represented the<br />
team <strong>and</strong> their <strong>work</strong>.<br />
Third, illustrating the previous point, no one in the team but Ethan<br />
actually took any initiative towards the PLENTI community. It was<br />
not until the lead programmers decided that something must be done<br />
about the lack <strong>of</strong> PLENTI skills <strong>and</strong> usage in the project, that<br />
effective interaction happened. In our case, it took a programmer not<br />
only to undertake the role as broker, but also to underst<strong>and</strong> that<br />
someone needed to take that role.<br />
6.1.3 <strong>The</strong> credibility gained through OSS participation<br />
Being acknowledged as a participant in a developer community is<br />
likely to add to the self consciousness <strong>of</strong> the participant. When a<br />
contribution is met by a response in a forum, the community<br />
membership <strong>of</strong> the contributor is acknowledged. When a<br />
contribution is openly praised in the forum, the skills <strong>of</strong> the<br />
participant are implicitly valued, <strong>and</strong> his credibility in the forum is<br />
maintained or strengthened. If the participant is relatively<br />
inexperienced <strong>and</strong> do not already participate actively on several<br />
related arenas, the newly earned credibility is likely to be a source<br />
<strong>of</strong> pride <strong>and</strong> increased self confidence. Also, the credibility won can<br />
be demonstrated to others outside the community, adding to<br />
credibility <strong>and</strong> strengthening authority there as well. From our case,<br />
we draw some illustrative points:<br />
In Exhibit 4, the last response from Bernhard to Ethan includes the<br />
following: “Wonderful! Several people have asked for it in the past,<br />
it would be a very welcome contribution to the project.” <strong>The</strong><br />
contribution <strong>of</strong> Ethan <strong>and</strong> the team is thus publicly acknowledged as<br />
something relevant <strong>and</strong> welcome. <strong>The</strong> positive response was<br />
referred to by Ethan on later occasions, e.g. in customer meetings.<br />
In team-internal conversation, Ethan on some occasions talked <strong>of</strong><br />
Bernhard almost as if he was a buddy, using his first name. Ethan<br />
also pointed out to the team how quick Bernard was at providing<br />
them with response, <strong>and</strong> <strong>of</strong>ten outside <strong>of</strong> normal <strong>work</strong> hours.<br />
Ethan’s pride in the engagement <strong>of</strong> the team with the PLENTI<br />
developer community was very visible whenever the project was<br />
accounted for in meetings with supervisor or customer in the last<br />
part <strong>of</strong> the project. On one h<strong>and</strong>, referring to ongoing successful<br />
interaction with the PLENTI community served as a way <strong>of</strong><br />
ensuring project stakeholders that the project was in fact<br />
progressing. On the other h<strong>and</strong>, it may be seen as a way <strong>of</strong> assuring<br />
the stakeholders that the team were competent, doing development<br />
<strong>work</strong> <strong>of</strong> a st<strong>and</strong>ard acceptable to the frame<strong>work</strong> developer<br />
community.<br />
Finally, whereas the fourth phase <strong>of</strong> OSS interaction, contributing to<br />
the frame<strong>work</strong>, might not have been strictly necessary for the team’s<br />
development task, it might have benefited the development task<br />
indirectly: Ethan’s positive attitude towards contributing to the<br />
PLENTI development might have inspired Bernhard to give better,<br />
faster or more elaborate response.<br />
6.2 OSS community participation aided by a<br />
broker: Benefits <strong>and</strong> pitfalls for SE projects<br />
Turning focus to what we see as benefits <strong>and</strong> pitfalls <strong>of</strong> having<br />
student SE teams acquire knowledge from OSS communities, we<br />
give a final account <strong>of</strong> the impact <strong>of</strong> OSS community involvement<br />
on the Anniva team, looking at their project more broadly.<br />
<strong>The</strong> team succeeded with their development task, producing a<br />
<strong>work</strong>ing system by active use <strong>of</strong> advice received from the OSS<br />
community. On the other h<strong>and</strong>, the project report suffered from a<br />
lack <strong>of</strong> attention from the programmers. <strong>The</strong> poor quality <strong>of</strong> the<br />
report was a main reason for the project receiving a B <strong>and</strong> not an A,<br />
according to course staff in the grading meeting.<br />
It is only a guess that the team in our case would have been no more<br />
concerned about the project report if they did not have to interact<br />
with an OSS community. <strong>The</strong> focus on programming (at the cost<br />
e.g. <strong>of</strong> developing models <strong>and</strong> documentation) might be seen as a<br />
consequence <strong>of</strong> a lack <strong>of</strong> experience <strong>and</strong> maturity in terms <strong>of</strong> project<br />
management, or as a legal choice <strong>of</strong> prioritizing what was to the<br />
team the essence <strong>of</strong> the project. In respect <strong>of</strong> the team postponing<br />
the <strong>work</strong> with design <strong>and</strong> architecture modeling, it is worth noting<br />
that the PLENTI frame<strong>work</strong> can be seen as an example <strong>of</strong> a third<br />
party s<strong>of</strong>tware component, unfamiliar to the team <strong>and</strong> poorly<br />
documented, which implies a realistic SE challenge <strong>of</strong> determining<br />
when to attempt to develop an overall design <strong>and</strong> when to start using<br />
the component [6].<br />
Communicating in the OSS forum appeared to be a gratifying<br />
experience for the broker. <strong>The</strong> team was vulnerable to his absence,<br />
but that problem never occurred. Also, our observations indicate<br />
that the broker shared his knowledge from the OSS forum with the<br />
rest <strong>of</strong> the team in an open <strong>and</strong> effective way. It is difficult to say if<br />
he spent too much <strong>of</strong> the project time in the forum, or if the team<br />
spent too much time on programming tasks based on the results <strong>of</strong><br />
the interaction. <strong>The</strong> fact that the interaction took place only in the<br />
last part <strong>of</strong> the project, indicates a limited use <strong>of</strong> project time. Also,<br />
the team’s interaction with the OSS community having to do with<br />
changes to PLENTI constituted only a small part <strong>of</strong> the interaction.<br />
Finally, these discussions were closely related to pertinent<br />
development tasks in the project.<br />
<strong>The</strong> fact that Ethan did not mention in the individual <strong>reflection</strong> note<br />
his role in the communication with the PLENTI community, might<br />
indicate that he did not consider it as something requiring a<br />
substantial part <strong>of</strong> his time or being very important to the project,<br />
but there may be other reasons why he did not mention his<br />
brokering role. As for the power <strong>of</strong> the broker, Owen (on reading a<br />
draft <strong>of</strong> this paper) expresses uncertainty about how much power<br />
Ethan gained through his gatekeeping role. Owen had not until now<br />
reflected much on the importance <strong>of</strong> the interaction with the<br />
PLENTI community. It may be the case that Owen <strong>and</strong> Ethan take<br />
their successful OSS community interaction for granted,<br />
underestimating the communication skills actually required.<br />
Finally, balancing our previous focus on the significance <strong>of</strong> OSS<br />
participation to the team, we should note that communicating with<br />
the PLENTI community was one out <strong>of</strong> several activities important<br />
to the project. <strong>The</strong> power structures in the team may be seen to have<br />
mirrored the complementary skills <strong>and</strong> interests <strong>of</strong> the two lead<br />
programmers, which ensured a certain balance in power between the<br />
two. Nevertheless, given our findings, our interpretation is that the<br />
team, <strong>and</strong> the broker in particular, were both challenged <strong>and</strong><br />
inspired by the fact that they were participants in, <strong>and</strong> contributors<br />
to the PLENTI OSS community <strong>and</strong> considered it an essential part<br />
<strong>of</strong> their project experience.<br />
Leaving the story <strong>of</strong> the auctioning system project, we now turn to<br />
what we see as more general benefits <strong>and</strong> pitfalls – in terms <strong>of</strong> a<br />
good project result - for student SE teams in need <strong>of</strong> interacting with<br />
OSS communities as part <strong>of</strong> their development <strong>work</strong>.<br />
798<br />
116
6.2.1 Benefits for SE student projects interacting with<br />
OSS communities<br />
In line with arguments from others’ research, we hold that OSS<br />
community participation is a realistic aspect <strong>of</strong> modern SE <strong>work</strong>.<br />
<strong>The</strong> management <strong>of</strong> such interaction should be regarded as highly<br />
relevant industry SE experience. If the team interacts with an OSS<br />
community, part <strong>of</strong> the experience is shared by the whole team.<br />
<strong>The</strong> broker in particular may learn a lot about how to achieve<br />
successful communication <strong>and</strong> acquire relevant knowledge in an<br />
OSS community. A broker whose authority <strong>and</strong> importance in the<br />
team is acknowledged by the team is likely to take pride in his<br />
position as a knowledge provider <strong>and</strong> thus be strongly motivated to<br />
contribute to the project result. If the broker further experiences that<br />
he earns credibility on the OSS arena, being recognized as a proper<br />
participant there, he might want to make his contribution more<br />
substantial <strong>and</strong> visible, which may again benefit his team.<br />
6.2.2 Pitfalls for SE student projects interacting with<br />
OSS communities<br />
Student SE teams may hesitate to embark on interaction over<br />
technically challenging issues in a type <strong>of</strong> community unknown to<br />
them, the students being in doubt about their own skills <strong>and</strong><br />
credibility. This may cause problematic project delays. Also,<br />
attempts to interact with an OSS community may turn out to be nonsuccessful<br />
if the team does not manage to convince the community<br />
that they represent ‘real’ users <strong>and</strong> development issues <strong>and</strong> thus<br />
qualify as potential contributors.<br />
To a team inexperienced in project management, interaction in an<br />
OSS community may pose great challenges because it involves an<br />
extra external stakeholder <strong>and</strong> may be perceived as even more out <strong>of</strong><br />
control than other programming related activity.<br />
OSS community participation can be used as an excuse to spend<br />
time on the wrong tasks from a project management point <strong>of</strong> view.<br />
Having an alternative arena for gratifying response <strong>and</strong><br />
acknowledged participation might lead a programmer to focus too<br />
narrowly on certain programming tasks related to that arena.<br />
Brokering may be performed in an inadequate way: A broker unable<br />
or unwilling to share relevant knowledge with the team, or mainly<br />
representing other interests than those <strong>of</strong> the team in the OSS<br />
community, st<strong>and</strong>s in the way <strong>of</strong> a good project result <strong>and</strong><br />
everyone’s <strong>learning</strong>. Further, there is a risk associated with having<br />
only one broker – both in case <strong>of</strong> inadequate brokering <strong>and</strong> in the<br />
case <strong>of</strong> sudden absence <strong>of</strong> the one broker.<br />
6.3 Implications for the pedagogy <strong>of</strong> SE<br />
project courses<br />
An implication <strong>of</strong> the <strong>work</strong> presented in this paper is that course<br />
staff responsible for SE project courses should pay particular heed<br />
to SE student projects requiring interaction with OSS communities<br />
for development-related knowledge acquisition. <strong>The</strong>se projects hold<br />
a potential for industry-relevant experience but at the same time<br />
hold some challenges. Course staff should be aware <strong>of</strong> benefits <strong>and</strong><br />
pitfalls inherent to this type <strong>of</strong> project. Further, awareness <strong>of</strong> the<br />
mechanism <strong>of</strong> brokering <strong>and</strong> the potentially influential role <strong>of</strong> the<br />
broker(s) in the projects may aid a supervisor in determining what<br />
advice to provide on project management <strong>and</strong> process issues.<br />
Part <strong>of</strong> the charm <strong>of</strong> industry projects is that they are very diverse.<br />
Having used our in-depth single case study to derive issues we<br />
believe to be generally relevant to student projects interacting with<br />
OSS communities as a resource in their development <strong>work</strong>, we<br />
would like to stress that the particular challenges <strong>of</strong> OSS<br />
participation in one project may be very different from those in<br />
another. Considerations over project management <strong>and</strong> supervision<br />
must always be made in the light <strong>of</strong> the particular characteristics <strong>of</strong><br />
the specific project. In this respect, the project supervisor will<br />
always benefit from having experience with the approaches <strong>and</strong><br />
domains relevant to the project. In the type <strong>of</strong> project <strong>of</strong> our current<br />
focus, some background within OSS development or research is<br />
likely to be an advantage to the supervisor.<br />
In a situation in which many <strong>of</strong> the projects in a course include OSS<br />
community participation, the overall organization <strong>of</strong> the course may<br />
be designed to reflect this particular focus. For instance, lectures<br />
may be given on how OSS communities <strong>work</strong>. Each project’s<br />
interaction with the OSS community may be more explicitly used as<br />
a <strong>learning</strong> resource, e.g. by being examined <strong>and</strong> discussed within<br />
<strong>and</strong> across teams. Further, the OSS community interaction <strong>of</strong> a team<br />
should be considered as a <strong>learning</strong> result (or documented part <strong>of</strong> the<br />
project process) <strong>and</strong> assessed <strong>and</strong> credited by course staff, e.g. in<br />
the grading <strong>of</strong> the projects.<br />
6.4 Limitations to the study<br />
Drawing on a single case, our findings may be seen as closely<br />
related to its particular characteristics. Still, we believe findings<br />
from our study on the role <strong>of</strong> the broker have general relevance to<br />
SE student projects interacting with OSS communities.<br />
An important characteristic <strong>of</strong> the OSS community <strong>of</strong> our case is its<br />
small size, which may be seen as a condition for the rapid entry <strong>of</strong><br />
the student team into OSS development contribution. Arguments<br />
about the empowerment <strong>of</strong> the broker however apply even if OSS<br />
community participation is restricted to the user role.<br />
Finally, our research explored a student team. Our findings may<br />
have some relevance to teams <strong>of</strong> SE pr<strong>of</strong>essionals, but differences<br />
between these categories <strong>of</strong> SE teams should be considered. <strong>The</strong>se<br />
include the level <strong>of</strong> competence in project management in the team,<br />
<strong>and</strong> the formalized <strong>learning</strong> goals <strong>of</strong> a project course which are not<br />
found in industry <strong>and</strong> which impact on how we wish to guide the SE<br />
student teams through project supervision.<br />
7. CONCLUSION<br />
In this study, we set out to gain insights on mechanisms enabling a<br />
SE student team to successfully interact with an open source<br />
community <strong>and</strong> on possible pitfalls <strong>and</strong> benefits <strong>of</strong> the engagement<br />
in terms <strong>of</strong> project success. We have done so by pointing to the role<br />
<strong>of</strong> the broker <strong>and</strong> how this role implies increased authority <strong>and</strong><br />
influence in the team for several reasons <strong>and</strong> with several possible<br />
consequences, positive <strong>and</strong> negative.<br />
We hope that our contribution will inspire SE course staff <strong>and</strong> others<br />
involved with SE student projects to build on our <strong>work</strong> with<br />
reference to their own experience. Continued focus on <strong>work</strong> <strong>and</strong><br />
<strong>learning</strong> in SE student projects involving OSS community<br />
participation should result in empirically based research<br />
contributions for us all to share, with the aim <strong>of</strong> improving current<br />
pedagogical practices.<br />
Further <strong>work</strong> may look at the implications <strong>of</strong> the type <strong>and</strong> sizes <strong>of</strong><br />
the OSS communities for student teams’ involvement there. Also,<br />
the relevance <strong>of</strong> OSS participation in student SE projects to similar<br />
799<br />
117
situations in real-life industry projects – <strong>and</strong> vice versa - are topics<br />
worth pursuing.<br />
8. ACKNOWLEDGMENTS<br />
<strong>The</strong> author wishes to thank Glenn Munkvold, John Krogstie, Owen<br />
<strong>and</strong> Ethan for feedback on draft versions <strong>of</strong> this paper. Monica<br />
Divitini <strong>and</strong> Thomas Østerlie provided valuable advice.<br />
9. REFERENCES<br />
[1] B. Krogstie <strong>and</strong> B. Bygstad, "Cross-Community Collaboration<br />
<strong>and</strong> Learning in Customer-Driven S<strong>of</strong>tware Engineering<br />
Student Projects," presented at CSEE&T, Dublin, 2006.<br />
[2] L. v. d. Duim, A. Jesper, <strong>and</strong> M. Sinnema, "Good practices for<br />
Educational S<strong>of</strong>tware Engineering Projects," presented at<br />
ICSE'07, Minneapolis, USA, 2007.<br />
[3] B. Bygstad, B. Krogstie, <strong>and</strong> T.-M. Grønli, "Scaffolding<br />
Project Based Learning with the Rational Unified Process.<br />
Experience from 5 years <strong>of</strong> Student Projects in S<strong>of</strong>tware<br />
Engineering," presented at NOKOBIT, Molde, 2006.<br />
[4] D. Wood, J. Bruner, <strong>and</strong> G. Ross, "<strong>The</strong> role <strong>of</strong> tutoring in<br />
problem solving " Journal <strong>of</strong> Child Psychology <strong>and</strong> Psychiatry,<br />
vol. 17, pp. 89-100, 1976.<br />
[5] P. H. Carstensen <strong>and</strong> K. Schmidt, "<strong>Computer</strong> Supported<br />
Cooperative Work: New Challenges to Systems Design," in<br />
H<strong>and</strong>book <strong>of</strong> Human Factors/Ergonomics, K. Itoh, Ed. Tokyo:<br />
Asakura Publishing, 2002 (1999).<br />
[6] J. Li, R. Conradi, C. Bunse, M. Torchiano, O. P. N. Slyngstad,<br />
<strong>and</strong> M. Morisio, "Development with Off-<strong>The</strong>-Shelf<br />
Components: 10 Facts," IEEE S<strong>of</strong>tware, 2007.<br />
[7] M. Riel <strong>and</strong> L. Polin, "Online Learning Communities.<br />
Common Ground <strong>and</strong> Critical Differences in Designing<br />
Technical Environments," in Designing for Virtual<br />
Communities in the Service <strong>of</strong> Learning, S. A. K. Barab,<br />
Rob; Gray, James H., Ed. Cambridge: Cambridge University<br />
Press, 2004, pp. 16-50.<br />
[8] E. Wenger, Communities <strong>of</strong> Practice. Learning, Meaning,<br />
<strong>and</strong> Identity: Cambridge University Press, 1998.<br />
[9] C. McDowell <strong>and</strong> L. Werner, "<strong>The</strong> Impact <strong>of</strong> Pair<br />
Programming on Student Performance, Perception <strong>and</strong><br />
Persistence," presented at ICSE'03, Portl<strong>and</strong>, Oregon, USA,<br />
2003.<br />
[10] J. Lave <strong>and</strong> E. Wenger, Situated Learning. Legitimate<br />
peripheral participation. Cambridge: University <strong>of</strong><br />
Cambridge Press., 1991.<br />
[11] M. Huysman, C. Steinfeld, C.-Y. Jang, K. David, M. Huis in<br />
't Veld, J. Poot, <strong>and</strong> I. Mulder, "Virtual Teams <strong>and</strong> the<br />
Appropriation <strong>of</strong> Communication Technology: Exploring the<br />
Concept <strong>of</strong> Media Stickiness," <strong>Computer</strong> Supported<br />
Cooperative Work, vol. 12, pp. 411-436, 2003.<br />
[12] C. Gutwin, R. Penner, <strong>and</strong> K. Schneider, "Knowledge<br />
sharing in s<strong>of</strong>tware engineering: Group awareness in<br />
distributed s<strong>of</strong>tware development " presented at CSCW'04,<br />
Chicago, Illinois, USA., 2004.<br />
[13] M. Bergquist <strong>and</strong> J. Ljungberg, "<strong>The</strong> power <strong>of</strong> gifts:<br />
organizing social relationships in open source communities,"<br />
Information Systems Journal, vol. 11, pp. 305-320, 2001.<br />
[14] E. Wenger, "Communities <strong>of</strong> practice <strong>and</strong> social <strong>learning</strong><br />
systems," Organization, vol. 7, pp. 225-246, 2000.<br />
[15] S. L. Star <strong>and</strong> J. R. Griesemer, "Institutional Ecology,<br />
'Translations' <strong>and</strong> Boundary Objects: Amateurs <strong>and</strong><br />
Pr<strong>of</strong>essionals in Berkley's Museum <strong>of</strong> Vertebrate Zoology,<br />
1907-39," Social Studies <strong>of</strong> Science, vol. 19, pp. 387-420,<br />
1989.<br />
[16] W. Sacchi, J. Feller, B. Fitzgerald, S. Hissam, <strong>and</strong> K.<br />
Lakhani, "Underst<strong>and</strong>ing Free/Open Source S<strong>of</strong>tware<br />
Development Processes," S<strong>of</strong>tware Process Improvement <strong>and</strong><br />
Practice, vol. 11, pp. 95-105, 2006.<br />
[17] K. Crowston <strong>and</strong> J. Howison, "<strong>The</strong> social structure <strong>of</strong> free<br />
<strong>and</strong> open source s<strong>of</strong>tware development," First Monday, vol.<br />
10, 2005.<br />
[18] A. Mockus, R. T. Fielding, <strong>and</strong> J. D. Herbsleb, "Two Case<br />
Studies <strong>of</strong> Open Source S<strong>of</strong>tware Development: Apache <strong>and</strong><br />
Mozilla," ACM Transactions on S<strong>of</strong>tware Engineering <strong>and</strong><br />
Methodology, vol. 11, pp. 309-346, 2002.<br />
[19] H. J. C. Ellis, R. A. Morelli, T. R. deLanerolle, <strong>and</strong> G. W.<br />
Hislop, "Holistic S<strong>of</strong>tware Engineering Education Based on a<br />
Humanitarian Open Source Project," presented at CSEE&T,<br />
Dublin, Irel<strong>and</strong> 2007.<br />
[20] K. Toth, "Experiences with Open Source S<strong>of</strong>tware<br />
Engineering Tools," IEEE S<strong>of</strong>tware, 2006.<br />
[21] D. Spinellis, "Open Source <strong>and</strong>Pr<strong>of</strong>essional Advancement,"<br />
IEEE S<strong>of</strong>tware, 2006.<br />
[22] L. Jaccheri <strong>and</strong> T. Østerlie, "Open Source S<strong>of</strong>tware: A<br />
Source <strong>of</strong> Possibilities for S<strong>of</strong>tware Engineering Education<br />
<strong>and</strong> Empirical S<strong>of</strong>tware Engineering," presented at First<br />
International Workshop on Emerging Trends in FLOSS<br />
Research <strong>and</strong> Development (FLOSS'07), 2007.<br />
[23] R. K. Yin, Case Study Research. Design <strong>and</strong> Methods. Third<br />
Edition., vol. 5: SAGE Publications, 2003.<br />
[24] H. K. Klein <strong>and</strong> M. M. Myers, "A Set <strong>of</strong> Principles for<br />
Conducting <strong>and</strong> Evaluating Interpretive Field Studies in<br />
Information Systems," MIS Quarterly, vol. 23, pp. 67-94,<br />
1999.<br />
800<br />
118
Research paper P3<br />
Title:<br />
Do’s <strong>and</strong> Don’ts <strong>of</strong> Instant Messaging in Students Project Work<br />
Authors:<br />
Birgit Krogstie<br />
Published in:<br />
Proceedings <strong>of</strong> NOKOBIT 2009<br />
Pages:<br />
14<br />
Complete reference:<br />
Krogstie, B. R. (2009). Do's <strong>and</strong> dont's <strong>of</strong> instant messaging in students' project <strong>work</strong>.<br />
NOKOBIT 2009, Trondheim, Norway, 23-25 Nov. Tapir.<br />
119
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
120
DO‟S AND DON‟TS OF INSTANT MESSAGING IN STUDENTS‟<br />
PROJECT WORK<br />
Birgit Krogstie, IDI, NTNU, birgitkr@idi.ntnu.no<br />
Abstract<br />
Instant messaging is a type <strong>of</strong> lightweight collaboration technology heavily used among students in<br />
higher education for social interaction <strong>and</strong> for school <strong>work</strong>. This paper sheds light on students’ use <strong>of</strong><br />
instant messaging to support collaboration in s<strong>of</strong>tware engineering student projects. We draw on several<br />
years’ <strong>of</strong> research on the use <strong>of</strong> collaboration technology in such projects <strong>and</strong> present illustrative<br />
examples <strong>of</strong> how instant messaging is used within the teams <strong>and</strong> for communication with other<br />
stakeholders, e.g. project supervisor <strong>and</strong> customer. We find that the use <strong>of</strong> instant messaging in the<br />
projects is generally successful, but in some cases it is inadequate, which may severely harm the project<br />
outcome. Based on our findings, we discuss implications for the organization <strong>and</strong> supervision <strong>of</strong> SE<br />
student projects.<br />
Keywords: instant messaging, synchronous chat, student projects, s<strong>of</strong>tware engineering student projects.<br />
1 INTRODUCTION<br />
“But in the case I have a technical question, <strong>and</strong> want to ask somebody in the group, I<br />
first look at MSN.” (Member <strong>of</strong> a student s<strong>of</strong>tware engineering project team)<br />
Observing s<strong>of</strong>tware engineering (SE) students in a project meeting or in the computer lab, you might<br />
notice chat windows frequently popping up on the laptop <strong>and</strong> terminal screens. Different actions are taken<br />
by the students in response to the pop-ups. Some windows are ignored (at least for the present), some are<br />
discreetly attended to with a brief textual exchange, <strong>and</strong> some (if the ongoing activity in the room allows<br />
it) appear to call for the immediate attention <strong>of</strong> the neighbour or all the students in the room. Some<br />
messages appear to be project or otherwise school related, others belong to the private sphere in which a<br />
social net<strong>work</strong> appears to be incessantly maintained mainly through this channel.<br />
What is going on, is instant messaging. Instant messaging tools are easily available, lightweight<br />
collaboration tools that provide the user with the opportunity to communicate with one or more contacts<br />
through textual chat. This channel <strong>of</strong> communication is constantly open <strong>and</strong> present in the students‟ <strong>work</strong><br />
environment; mostly in the background <strong>and</strong> sometimes in the foreground. From the viewpoint <strong>of</strong> course<br />
staff in a project course, it is highly interesting to know whether, <strong>and</strong> in what way, the use <strong>of</strong> instant<br />
messaging is supporting, or perhaps getting in the way <strong>of</strong>, the intended <strong>learning</strong>.<br />
SE student projects is a case <strong>of</strong> project-based <strong>learning</strong> (Blumenfeld, Soloway et al. 1991). This<br />
pedagogical approach is based on student projects that are central to the curriculum, focused on questions<br />
or problems that “drive” students to encounter the central concepts <strong>and</strong> principles <strong>of</strong> a discipline,<br />
involving students in a constructive investigation, student-driven to a significant degree, <strong>and</strong> realistic, not<br />
school-like (Thomas 2000). Through SE student projects, students are being exposed to challenges <strong>of</strong> SE<br />
<strong>work</strong> while receiving necessary guidance <strong>and</strong> support (Bygstad, Krogstie et al. 2006; Bygstad, Krogstie et<br />
al. 2009). Figure 1 shows a SE student team in various settings <strong>of</strong> everyday <strong>work</strong>.<br />
121<br />
13
Figure 1.<br />
A SE student team at <strong>work</strong><br />
SE is complex design <strong>work</strong> (Carstensen <strong>and</strong> Schmidt 2002 (1999)), <strong>and</strong> project teams have to manage<br />
project tasks under constraints such as deadlines, team members‟ competence <strong>and</strong> customers‟ changing<br />
product requirements. While some <strong>of</strong> the project challenges are technical, many are mainly about<br />
interpersonal issues. Collaboration with project stakeholders (e.g. customer, supervisor, technology<br />
providers) is frequently challenging to the teams (Krogstie <strong>and</strong> Bygstad 2007; Krogstie 2008).<br />
Stakeholder communication is an area in which industry points to a need for SE students to gain<br />
competence through experience in student projects (McMillan 1999).<br />
<strong>The</strong> <strong>work</strong> in SE projects is supported by collaboration technology. In the case <strong>of</strong> many SE industry<br />
projects, <strong>and</strong> typically in student projects, lightweight collaboration tools are used. Lightweight<br />
collaboration tools can be acquired <strong>and</strong> taken into use at low cost (e.g. money, time to learn) for<br />
individuals <strong>and</strong> their organization. A lightweight collaboration tool typically provides a limited set <strong>of</strong><br />
features to support one aspect <strong>of</strong> collaborative <strong>work</strong> <strong>and</strong> may thus be relatively easily integrated into<br />
existing <strong>work</strong> processes (rather than imposing a certain process on the user). Many <strong>of</strong> the lightweight<br />
collaboration tools are associated with Web 2.0, e.g. wikis, discussion forums, <strong>and</strong> instant messaging.<br />
In this paper, we shed light on the use <strong>of</strong> instant messaging tools in SE student teams. Our main questions<br />
are: when IM is used to aid project <strong>work</strong>, for what purposes <strong>and</strong> in what way is IM used? And what are<br />
the implications for the organization <strong>of</strong> SE project courses? <strong>The</strong> findings reported in the paper originate in<br />
qualitative studies <strong>of</strong> three cohorts <strong>of</strong> SE project students, 2006, 2007 <strong>and</strong> 2008, in which we investigated<br />
the students‟ use <strong>of</strong> collaboration technology in their project <strong>work</strong>.<br />
<strong>The</strong> paper is organized as follows: In Section 2 we briefly describe the functionality <strong>of</strong> IM tools <strong>and</strong><br />
present related <strong>work</strong> on the use <strong>of</strong> instant messaging among students <strong>and</strong> in <strong>work</strong> life. In Section 3 we<br />
outline our research approach. Findings on the use <strong>of</strong> IM in the SE student projects <strong>of</strong> our study are<br />
presented in Section 4. In Section 5 we discuss the findings in light <strong>of</strong> implications for the the<br />
organization <strong>and</strong> supervision <strong>of</strong> SE student projects. Section 6 concludes the paper.<br />
2 BACKGROUND<br />
We start this section by briefly outlining the main functionality <strong>of</strong> IM tools. <strong>The</strong>re are many different IM<br />
platforms. <strong>The</strong>y include AOL‟s Instant Messenger <strong>and</strong> Micros<strong>of</strong>t‟s Windows Messenger (MSN). Key<br />
attributes <strong>of</strong> IM tools are (Garrett <strong>and</strong> Danziger 2008):<br />
<br />
<br />
<br />
Near-synchronous communication that can be initiated by either party<br />
Presence awareness showing if other users are connected <strong>and</strong>/or available<br />
Notifications <strong>of</strong> incoming communication, typically with pop-up windows<br />
Many tools support the exchange <strong>of</strong> speech <strong>and</strong> video as well, but in what follows we will focus on what<br />
can be considered the basic functionality: textual chat. This is the functionality used by the SE students in<br />
our case.<br />
When a user logs on to the IM service, a „buddy list‟ appears on the screen. <strong>The</strong> list shows the user‟s<br />
contacts <strong>and</strong> updated information on their status: whether they are logged on to the service, <strong>and</strong> any<br />
message they might have left indicating their availability <strong>and</strong>/or current activity. “Busy reading for exam”<br />
could for instance be such a message. If the user chooses to initiate a chat with a contact in the buddy list,<br />
122<br />
14
a chat window is opened. As soon as the user has written something <strong>and</strong> clicked „Enter‟, a chat window<br />
appears on the contact‟s screen. <strong>The</strong> chat may also include more than two participants.<br />
Once the user has written one or more lines <strong>of</strong> text in the chat window <strong>and</strong> clicks „Enter‟, the text will be<br />
sent to the other party/parties, who may choose to answer directly, in the same way. In this sense, instant<br />
messaging can be seen as a synchronous communication medium. On the other h<strong>and</strong>, instant messaging is<br />
„semi-asynchronous‟ in the sense that the message will remain in the recipient‟s open chat window <strong>and</strong><br />
may be answered at a later point, e.g. if the recipient is away for a while or if he wants to finish other<br />
tasks first. Further, many instant messaging services support <strong>of</strong>fline messaging, which means that a<br />
message may be sent to a contact who is <strong>of</strong>fline to be delivered when he logs on to the system. <strong>The</strong> user<br />
may choose to save the conversation in a log. Instant messaging can thus be seen as a „semi-persistent‟<br />
communication medium.<br />
Research shows that IM tools have generally come to play an important role in the life <strong>of</strong> young people.<br />
IM tools provide a means for developing <strong>and</strong> maintaining peer group membership <strong>and</strong> social identity<br />
(Grinter <strong>and</strong> Palen 2002; Grinter <strong>and</strong> Eldridge 2003; Lewis <strong>and</strong> Fabos 2005). Young people “meet on<br />
MSN”. <strong>The</strong>y are also conscious about how IM supports both social <strong>and</strong> <strong>work</strong>-related interaction <strong>and</strong> what<br />
features <strong>of</strong> IM are desirable for different purposes (Huang <strong>and</strong> Yen 2003). Young people‟s use <strong>of</strong> IM<br />
tools, <strong>and</strong> the links to other members <strong>of</strong> the social net<strong>work</strong> contained in the tools, reflect a „net<strong>work</strong>ed<br />
individualism‟ <strong>and</strong> participation in a net<strong>work</strong>ed society (Castells 2000). A study addressing American<br />
college students‟ perception <strong>of</strong> the differences between instant messaging <strong>and</strong> email (Lancaster, Yen et al.<br />
2007) concludes that IM is seen by the college students to be the best option in the context <strong>of</strong> personal<br />
<strong>and</strong> social relationships, IM being better for conveying emotions, building relationships <strong>and</strong> in ease <strong>of</strong><br />
use. From this research, we may infert that SE project students are likely to be skilled users <strong>of</strong> instant<br />
messaging, utilizing the opportunities that have opened up by the popularity <strong>and</strong> extensive use <strong>of</strong> these<br />
tools, <strong>and</strong> being aware <strong>of</strong> some <strong>of</strong> their limitations. However, we cannot assume that the students are<br />
equally familiar with the use <strong>of</strong> IM in a context <strong>of</strong> <strong>work</strong>, for which other considerations might apply.<br />
In the <strong>work</strong>place, IM has become a mainstream informal collaboration tool (Cherry 2002; Stephens<br />
2008) <strong>and</strong> has been found to support a number <strong>of</strong> different communication tasks as well as the process <strong>of</strong><br />
enabling information exchange, i.e. reaching out to others to initiate communication (Nardi, Whittaker et<br />
al. 2000). Among the types <strong>of</strong> IM use is complex <strong>work</strong> discussions (Isaacs, Walendowski et al. 2002).<br />
Research on open source development (OSD) has shown that a purely textual medium (e.g. IRC) which is<br />
not integrated with the development tools may be sufficient for effective <strong>work</strong>-related communication<br />
between developers (Gutwin, Penner et al. 2004). When our SE project students use instant messaging,<br />
they are, in other words, getting experience with collaboration technology in industry use.<br />
Whereas IM is in itself interruptive, a recent study found that overall it helps people manage their<br />
interruptions rather than increasing the overall amount <strong>of</strong> <strong>work</strong> disruption (Garrett <strong>and</strong> Danziger 2008).<br />
This has to do with the ways IM allows users to manage availability. One way <strong>of</strong> doing this is to flag it in<br />
the buddy list (e.g. “<strong>work</strong>ing hard, will not answer”). Further, it is generally socially acceptable to<br />
postpone IM response, because non-response means that you might be away from the computer. This is<br />
what Nardi (Nardi, Whittaker et al. 2000) calls “plausible deniability”. <strong>The</strong> new communication patterns<br />
afforded by IM also support interruption management (Garrett <strong>and</strong> Danziger 2008): quick answers can be<br />
obtained without expectation <strong>of</strong> longer conversation, <strong>and</strong> a chat window can be kept open among the<br />
communicating parties for as long as desired, allowing for queries on at-need basis with the expectation<br />
<strong>of</strong> having an answer when it is convenient for the other party. <strong>The</strong> findings on IM tools as tools for<br />
interruption management can be taken as an argument that in the context <strong>of</strong> project-based <strong>learning</strong> we<br />
should not be too concerned about the popping up <strong>of</strong> chats windows irrelevant to the project <strong>work</strong> at<br />
h<strong>and</strong>. <strong>The</strong> students are likely to be capable <strong>of</strong> managing their use <strong>of</strong> time, helped by the social norms <strong>and</strong><br />
sanctions within the team, <strong>and</strong> having to manage their time is part <strong>of</strong> the desired project <strong>work</strong> experience.<br />
Rather than focusing on the use <strong>of</strong> IM that is not relevant to project <strong>work</strong>, thus, in what follows we focus<br />
on the use <strong>of</strong> IM that is related to the project <strong>work</strong>.<br />
In the SE projects <strong>of</strong> our study, different collaboration tools were used to support different aspects <strong>of</strong><br />
<strong>work</strong>. While the specific selection <strong>and</strong> usage <strong>of</strong> tools differed among teams, there was a pattern (Krogstie<br />
2009): Lightweight project management tools (typically project wikis or issue tracking tools) were used<br />
123<br />
15
for managing team-internal task coordination <strong>and</strong> provide links to project documentation. Development<br />
tools were used to write, test <strong>and</strong> integrate source code. Storage <strong>and</strong> versioning <strong>of</strong> project artifacts were<br />
managed in a file versioning system that may or may not be integrated with the project management tool.<br />
Email was used for formal <strong>and</strong> documented team-internal <strong>and</strong> external communication. Project teams<br />
typically had their own mailing list. Internet sites were used to get necessary information about<br />
technology. This included FAQ lists <strong>and</strong> discussion forums. IM was used for informal messages within<br />
the team <strong>and</strong> to support synchronous, distributed <strong>work</strong>. Less frequently, it was used for communication<br />
with other stakeholders.<br />
In this paper, we will go into detail about the teams‟ use <strong>of</strong> instant messaging.<br />
3 RESEARCH APPROACH<br />
<strong>The</strong> paper is based on empirical research on SE student teams in the period 2006-2008 focusing on the<br />
teams‟ use <strong>of</strong> collaboration technology (Krogstie <strong>and</strong> Bygstad 2007; Krogstie 2008; Krogstie 2008;<br />
Krogstie 2009). <strong>The</strong> data have been collected from two project courses <strong>of</strong>fered in the last (6th) semester<br />
<strong>of</strong> undergraduate programs in IT at two different <strong>learning</strong> institutions (NITH <strong>and</strong> NTNU). In the courses,<br />
teams <strong>of</strong> 3-5 members develop s<strong>of</strong>tware for genuine customers. <strong>The</strong> main <strong>learning</strong> objective is to have<br />
students get experience with SE <strong>work</strong> in a team for a customer. Project deliveries include a s<strong>of</strong>tware<br />
product <strong>and</strong> a report. <strong>The</strong> teams choose which collaboration <strong>and</strong> development tools to use, depending on<br />
customer requirements, team members‟ prior experience, team members‟ wish to learn, <strong>and</strong> the<br />
availability <strong>of</strong> the technology. <strong>The</strong> NITH <strong>and</strong> NTNU project courses are considered similar enough for<br />
data on the teams‟ use <strong>of</strong> collaboration technology to be combined in our study.<br />
<strong>The</strong> research on instant messaging is based on triangulation <strong>of</strong> data sources. <strong>The</strong>y include 30-60 minutes‟<br />
interviews about the use <strong>of</strong> collaboration technology in the teams, performed with all 7 teams in the NITH<br />
course in 2006, all 9 teams in the NTNU course in 2007, <strong>and</strong> 10 out <strong>of</strong> the 11 teams in 2008. <strong>The</strong>y also<br />
include longitudinal observation <strong>of</strong> two NTNU teams (one in 2007 <strong>and</strong> one in 2008) over the entire<br />
semester, <strong>and</strong> project deliveries <strong>and</strong> customer evaluation for all teams in the NITH course in 2006 <strong>and</strong> the<br />
NTNU course in 2006-2008. Further, some NTNU teams in 2006-2008 have been interviewed after their<br />
project based on interesting characteristics <strong>of</strong> their use <strong>of</strong> collaboration technology. In some cases, we<br />
have asked teams for IM logs. Overall, the data from NITH have been used in combination with the<br />
NTNU data to identify general patterns <strong>of</strong> collaboration technology use in the SE student teams. In-depth<br />
study <strong>of</strong> IM use in single teams, with access to IM logs, has been conducted only with data from the<br />
NTNU course.<br />
<strong>The</strong> researchers‟ role as course staff at both <strong>learning</strong> institutions provided a background for identification<br />
<strong>of</strong> interesting data <strong>and</strong> for their interpretation.<br />
During data collection <strong>and</strong> analysis, some themes <strong>and</strong> concepts have been pursued from the start (e.g.<br />
Instant Messaging), <strong>and</strong> some have emerged through the analysis. Summaries <strong>and</strong> transcripts have been<br />
produced at need, <strong>and</strong> translation to English performed as necessary for publication.<br />
<strong>The</strong> research for the paper can be considered mainly as an interpretive case study in which data collection<br />
<strong>and</strong> analysis have been guided by a constructivist view <strong>of</strong> <strong>learning</strong> <strong>and</strong> guidelines for interpretive field<br />
studies (Klein <strong>and</strong> Myers 1999). However, in the paper, rather than focusing in detail on “the complexity<br />
<strong>of</strong> human sense making as the situation emerges” <strong>and</strong> underst<strong>and</strong>ing “phenomena through the meanings<br />
that people assign to them” (Klein <strong>and</strong> Myers 1999), which typically implies an in-depth elaboration <strong>of</strong><br />
single cases, we wanted to present several examples to illustrate how instant messaging is used in<br />
different ways in the projects <strong>and</strong> link the examples to general patterns <strong>of</strong> IM use seen from the<br />
interviews. By drawing some inferences across the cases we thus approach a positivist type <strong>of</strong> case study<br />
research (Yin 2003).<br />
4 FINDINGS<br />
Findings are presented in the following way: First, team-internal use <strong>of</strong> IM is addressed <strong>and</strong> next, we turn<br />
to the use <strong>of</strong> IM in collaboration with project stakeholders. In our presentation we have given priority to<br />
124<br />
16
three examples in the form <strong>of</strong> excerpts from real IM logs, to provide the reader with a flavor <strong>of</strong> the actual<br />
IM usage in the teams. While our research approach does not enable us to quantify the representativeness<br />
<strong>of</strong> the examples, we address more informally in the below sections what can be considered typical <strong>and</strong><br />
untypical.<br />
4.1 Team-internal use <strong>of</strong> instant messaging<br />
A type <strong>of</strong> IM usage very frequently seen in the teams, is brief coordination <strong>and</strong> information sharing<br />
within the team, synchronous <strong>and</strong> asynchronous. IM is a channel constantly open to the other team<br />
members <strong>and</strong> used as a convenient way to get an indication about others‟ presence (in the buddy list) or<br />
sharing information even when the other team member is present <strong>and</strong> sitting nearby. Frequently, small<br />
amounts <strong>of</strong> <strong>work</strong>space contents are shared by pasting it directly into a chat window, which may be very<br />
effective to provide context for a question or with concrete information needed for a specific task. In the<br />
world <strong>of</strong> s<strong>of</strong>tware engineering, exact syntax is <strong>of</strong>ten essential to make things <strong>work</strong>, which makes „cut <strong>and</strong><br />
paste‟ very useful.<br />
In general, from our observations, the use <strong>of</strong> IM for these brief exchanges appears to be well integrated<br />
into the students‟ way <strong>of</strong> communicating with their peers <strong>and</strong> presents few problems to the projects. In<br />
line with the <strong>work</strong> on interruption management (Garrett <strong>and</strong> Danziger 2008) we note that the students<br />
cope with always being „on‟ <strong>and</strong> available.<br />
Team-internal IM use is however not limited to brief exchanges. Example 1 shows an excerpt from a<br />
conversation between Dave <strong>and</strong> Chuck, two out <strong>of</strong> the three members in a 2007 team, <strong>and</strong> the ones most<br />
involved in the team‟s development <strong>work</strong>. Dave <strong>and</strong> Chuck were generally <strong>work</strong>ing very well together,<br />
for several reasons preferring a distributed <strong>work</strong> arrangement <strong>and</strong> using IM to discuss ongoing <strong>work</strong>.<br />
<strong>The</strong>y had access to a shared server <strong>and</strong> <strong>work</strong>ed on different, but integrated parts <strong>of</strong> the system. With this<br />
partially shared <strong>work</strong>space, Dave <strong>and</strong> Chuck could immediately see some <strong>of</strong> the results <strong>of</strong> the other‟s<br />
<strong>work</strong>, but not all. Results could however be communicated through the IM chat window.<br />
In the excerpt shown in Example 1, the project is approaching deadline, <strong>and</strong> Dave <strong>and</strong> Chuck are <strong>work</strong>ing<br />
from their respective homes late at night. <strong>The</strong> conversation takes place between 03:57:58 <strong>and</strong> 04:10:53<br />
a.m. <strong>The</strong> overarching task is debugging: identifying <strong>and</strong> correcting errors in the code.<br />
<strong>The</strong> excerpt contains several references to the development <strong>work</strong> going on in parallel, both the current<br />
events in the shared <strong>work</strong>space (e.g. “pop, a notepad”, Dave referring to a text-editing window appearing<br />
on his screen) <strong>and</strong> ongoing tasks for which the two students have distributed responsibility (e.g. Dave:<br />
“but what about conditionals? do they run correctly”. Chuck: “don‟t think move runs correctly”. Dave:<br />
“ok”). A shared underst<strong>and</strong>ing <strong>of</strong> the current state <strong>of</strong> <strong>work</strong> is thus constantly maintained <strong>and</strong> re-created.<br />
New tasks are identified, negotiated <strong>and</strong> assigned. (Dave: “can you debug it?”. Chuck:”I‟ll just have a<br />
look at the installer first”. Dave:”ok”).<br />
Humor is extensively used in the conversation, see for instance the emoticons like “:OOO” <strong>and</strong> the<br />
outburst “give me that much <strong>work</strong> because <strong>of</strong> nothing!”. <strong>The</strong> generally positive tone <strong>and</strong> close<br />
relationship between the two students permits remarks like the last one to be given without resulting<br />
tension. Also, the tone <strong>of</strong> the conversation should be seen in light <strong>of</strong> the ongoing final project spurt <strong>and</strong><br />
fact that the <strong>work</strong> <strong>and</strong> the conversation takes place in the middle <strong>of</strong> the night.<br />
Incidentally, the excerpt in Example 1 does not include the sharing <strong>of</strong> textual elements like pieces <strong>of</strong><br />
source code or error messages, but the whole chat from which the excerpt is taken contains much <strong>of</strong> this<br />
type <strong>of</strong> material.<br />
<strong>The</strong> entire log shows that the chat window was kept open for about 29 hours. This includes breaks when<br />
the team members went to get some sleep, food, c<strong>of</strong>fee, or take a shower. <strong>The</strong> log contains comments<br />
addressing these activities.<br />
125<br />
17
Example 1.<br />
Excerpt from an IM log illustrating conversation over ongoing, distributed debugging<br />
<strong>work</strong> among the students Dave <strong>and</strong> Chuck. Translated from Norwegian by the author.<br />
In sum, the open IM window substituting face-to-face communication in cooperative development <strong>work</strong><br />
among the two students provides both <strong>work</strong>space awareness (Gutwin <strong>and</strong> Greenberg 2002) <strong>and</strong> social<br />
awareness to the parties. As a note, we may add that Chuck <strong>and</strong> Dave‟s team delivered an outst<strong>and</strong>ing<br />
project result.<br />
4.2 Use <strong>of</strong> instant messaging in collaboration with external stakeholders<br />
Generally, we see that when team <strong>and</strong> stakeholder communicate over IM, it is successful for quick <strong>and</strong><br />
informal exchanges. Example 2 shows a case <strong>of</strong> this in the form <strong>of</strong> an IM exchange between a team <strong>and</strong> a<br />
supervisor. <strong>The</strong> chat takes place during <strong>work</strong> hours, lasts for 1.5 minutes <strong>and</strong> addresses an upcoming<br />
project delivery. <strong>The</strong> conversation takes up on previous conversation on what is to be h<strong>and</strong>ed in by the<br />
team. Besides answering Michael‟s initial question whether the activity plan for the last two-week period<br />
is to be delivered, Thomas (the supervisor) makes sure that the team remembers to h<strong>and</strong> in both the<br />
original plan <strong>and</strong> a revised version showing the actual <strong>work</strong> done. Michael informs that the team is going<br />
to revise the delivery document internally <strong>and</strong> send the final version during the same evening, which<br />
Thomas accepts. <strong>The</strong> exchange can be seen as efficient, friendly, informal <strong>and</strong> task-oriented, <strong>and</strong> reveals<br />
an accommodating attitude on part <strong>of</strong> the supervisor.<br />
126<br />
18
Example 2.<br />
Excerpt from an IM log showing a brief exchange between a project manager (Michael)<br />
in a student SE team <strong>and</strong> their supervisor (Thomas). Translated from Norwegian by the<br />
author.<br />
<strong>The</strong> extent to which instant messaging is used for team-supervisor communication in the projects <strong>of</strong> our<br />
study depends on teams‟ <strong>and</strong> supervisors‟ preferences. Supervisors who communicate with project teams<br />
over IM, typically communicate with all their teams on this channel. When (most) supervisors do not use<br />
IM for communication with the teams, it may be because they are not heavy IM users or because they<br />
restrict availability to other channels. In general, the teams appear not to expect from the beginning that<br />
the supervisor be available through IM. <strong>The</strong> main thing for the students is that there is an agreement on<br />
how communication should take place <strong>and</strong> the supervisor is available <strong>and</strong> responds as agreed, whether<br />
face-to-face, over IM or email. In cases where communication should be documented, as is <strong>of</strong>ten the case,<br />
email <strong>and</strong> meetings with minutes are favored by teams <strong>and</strong> supervisors.<br />
For communication between teams <strong>and</strong> customers, email <strong>and</strong> face- to-face meetings are the most<br />
frequently used channel. In cases <strong>of</strong> remote customer, telephone meetings <strong>and</strong> video conferencing are also<br />
used.<br />
However, particularly in the case <strong>of</strong> remote customers, instant messaging is sometimes being used for<br />
informal team-customer communication, typically over technical issues. An example <strong>of</strong> the latter is a<br />
2008 team whose customer was located in another city. <strong>The</strong> following excerpt is from the team‟s<br />
<strong>reflection</strong> note addresses the issue <strong>of</strong> which type <strong>of</strong> cooperation technology was used for which purpose:<br />
“MSN: <strong>The</strong> group created a separate MSN account that was used for customer contact. <strong>The</strong><br />
customer was always available on MSN when he was at <strong>work</strong>, which was very useful<br />
whenever we had questions that needed to be answered quickly.”<br />
<strong>The</strong> customer <strong>of</strong> this team expresses similar satisfaction in his final evaluation <strong>of</strong> the project, commenting<br />
that<br />
“We have had good communication over email <strong>and</strong> MSN. <strong>The</strong> team asked for input whenever<br />
they were uncertain about something.”<br />
A 2007 team also had a customer who was located in another city for large parts <strong>of</strong> the project<br />
period, occasionally visiting for face-to-face meetings with the team. While remote, he used instant<br />
messaging actively in communication with the team. In their common <strong>reflection</strong> note, the team writes:<br />
“MSN has been the instant messenger application <strong>of</strong> choice for all group members, <strong>and</strong><br />
enabled us to easily stay in touch with each other <strong>and</strong> the customer at all times <strong>of</strong> the day<br />
(<strong>and</strong> night, for that matter).”<br />
127<br />
19
In his formal evaluation <strong>of</strong> the process after the project, the customer particularly expresses appreciation<br />
<strong>of</strong> the „continuous communication‟ with the team during the period when he was in another city.<br />
In both <strong>of</strong> the above cases the customer took a strong personal interest in the development process <strong>and</strong><br />
appreciated being involved with technical issues <strong>and</strong> have them resolved quickly. Formal meetings were<br />
however not conducted over IM.<br />
A team in the 2006 cohort did use IM for formal meetings with their remote customer. While this is not<br />
typical <strong>of</strong> our projects, we will go into some detail on the case because it demonstrates a communication<br />
breakdown which might in part be due to the inadequacy <strong>of</strong> IM in the particular coollaboration setting.<br />
<strong>The</strong> project group in question had taken on the task <strong>of</strong> making a game for a collaborative virtual<br />
environment. <strong>The</strong> customer role was held by a team <strong>of</strong> two researchers, one at the students‟ university <strong>and</strong><br />
one in an academic institution on the other side <strong>of</strong> the globe. <strong>The</strong> team never met the remote customer<br />
face-to-face. Meetings between him <strong>and</strong> the group were conducted via IM. Being the one proposing the<br />
use <strong>of</strong> IM for meetings, the customer was willing to have meetings at late hours to cope with the time<br />
difference. As the project process was being subject to formal evaluation, the project group wanted to<br />
document their communication with the customer. <strong>The</strong>y therefore chose, with their customer‟s consent, to<br />
log the meetings conducted via IM. Some weeks into the project, the group was frustrated because <strong>of</strong><br />
what they conceived to be misunderst<strong>and</strong>ings <strong>and</strong> vagueness related to needs, requirements,<br />
responsibilities <strong>and</strong> expectations in the project. Deadline was steadily approaching without the necessary<br />
progress <strong>of</strong> <strong>work</strong>. Example 3 shows an excerpt from a meeting over IM conducted at this rather critical<br />
point in the project. We have called the remote customer Johnatan/Johnny, <strong>and</strong> his Norwegian partner<br />
Anja. Frode <strong>and</strong> Per are two <strong>of</strong> the four students in the project group. During the meeting, the students are<br />
collocated in the PC lab. Anja is on another location in the students‟ city whereas Johnny is on the other<br />
side <strong>of</strong> the globe. <strong>The</strong> communication is taking place in English, Johnny‟s mother tongue. We lack<br />
timestamps for the comments in the log, but the conversation shown in the excerpt should be understood<br />
as continuous <strong>and</strong> without major pauses.<br />
From Example 3 it can be noted that highly critical questions in the project are at stake: what are the<br />
requirements, who is doing the planning <strong>and</strong> resource allocation (the group feels it is their job <strong>and</strong> that the<br />
customer is interfering), <strong>and</strong> access to important resources for the <strong>work</strong> (a particular server). In the heat <strong>of</strong><br />
the conversation, Frode, who was in practice the project manager, starts comm<strong>and</strong>ing the remote<br />
customer in a mix <strong>of</strong> anger <strong>and</strong> desperation: “Read the storyline document! Get us access to the server”.<br />
Not surprisingly, the remote customer simply leaves the conversation at that point, <strong>and</strong> communication<br />
breaks down. <strong>The</strong> team expresses uneasiness with the situation. <strong>The</strong> local customer, Anja, stays in the<br />
conversation <strong>and</strong> tries to moderate <strong>and</strong> explain the situation. Frode goes on to explain what he sees as the<br />
major problems.<br />
128<br />
20
Example 3.<br />
Excerpt from an IM log showing a communication breakdown between a student SE team<br />
<strong>and</strong> their remote customer. Originally in English.<br />
5 DISCUSSION.<br />
In this section, we discuss the findings <strong>and</strong> outline implications for the organizing <strong>of</strong> SE project courses.<br />
Research referred to in Section 2 indicated that young people are likely to be competent users <strong>of</strong> instant<br />
messaging tools. Our findings confirm that this is the case among the SE project students. Being a simple,<br />
lightweight tool already part <strong>of</strong> the students‟ personal infrastructure as a means for communication <strong>and</strong><br />
social net<strong>work</strong>ing, it appears that IM is easily incorporated into project based <strong>learning</strong> in the SE projects<br />
as a way <strong>of</strong> supporting informal communication; synchronous <strong>and</strong> asynchronous, distributed <strong>and</strong><br />
collocated. Further, IM provides a limited but easy way <strong>of</strong> sharing <strong>work</strong>space content by cut-<strong>and</strong>-paste<br />
(<strong>and</strong> even the exchange <strong>of</strong> files via shared folders). <strong>The</strong> fact that instant messaging is also in use among<br />
developers in industry SE projects stresses the value <strong>of</strong> using <strong>and</strong> developing students’ IM competence.<br />
From the assumption that underst<strong>and</strong>ing the students‟ <strong>work</strong> process is a prerequisite to providing good<br />
supervision <strong>of</strong> the process, we believe that increased insight about the current <strong>and</strong> potential use <strong>of</strong> IM<br />
tools to support <strong>work</strong> can help course staff provide students with guidance on some aspects <strong>of</strong> the<br />
students‟ <strong>work</strong> process. General findings on the use <strong>of</strong> instant messaging in student teams as presented in<br />
this paper are one source <strong>of</strong> knowledge about IM usage. Much more important in the specific case <strong>of</strong><br />
supervision are the particular practices <strong>of</strong> the team in question. As opposed to project activities being<br />
formally documented e.g. in email or deliverables, students‟ IM use is typically inaccessible to the<br />
supervisor. Leaving out the option that the supervisor be provided access to, <strong>and</strong> examine, the team‟s IM<br />
logs, the teams must provide their supervisor with information about their IM collaboration practices in<br />
order to get advice on these practices. Where problems <strong>of</strong> collaboration occur in the projects, it should be<br />
129<br />
21
considered whether current use <strong>of</strong> IM plays a role. Where IM is successfully supporting project <strong>work</strong>, it<br />
might still be worthwhile for the team to account for the tool usage, to increase their underst<strong>and</strong>ing <strong>of</strong><br />
their own <strong>work</strong> practice <strong>and</strong> to become aware <strong>of</strong> potential problems.<br />
Distributed project <strong>work</strong> can be successfully supported by instant messaging as a communication channel<br />
combined with shared <strong>work</strong>space access, as seen in Example 1. <strong>The</strong> example illustrated how the social<br />
<strong>and</strong> the task-oriented aspects <strong>of</strong> <strong>work</strong>ing together were both present in the communication taking place<br />
through the tool. Such an arrangement should be considered an option in projects that for some reason<br />
depends on <strong>work</strong> to be conducted in a distributed way (e.g. when one or more students have to partially<br />
<strong>work</strong> from home). It should be noted, however, that in our projects, the team members‟ <strong>work</strong> is partially<br />
collocated. <strong>The</strong> likelihood <strong>of</strong> success <strong>of</strong> IM chat as a substitute for face-to-face interaction may be<br />
different with students who do not know each other from face-to-face interaction, but we cannot draw<br />
conclusions about differences between IM use in virtual <strong>and</strong> partially collocated teams from our study.<br />
<strong>The</strong> practice <strong>of</strong> letting the student teams choose the collaboration technology for their projects is in line<br />
with the independence seen as important in project-based <strong>learning</strong>. Also, it can be seen as a way <strong>of</strong><br />
encouraging students to make use <strong>of</strong> their existing competence with certain tools, e.g. IM tools, <strong>and</strong><br />
consider how they should be applied in a new <strong>and</strong> challenging context (e.g. that <strong>of</strong> coordinating complex<br />
<strong>work</strong> tasks <strong>and</strong> collaborating with external stakeholders). Being allowed to choose the collaboration<br />
technology for their projects, we assume that most teams will choose to use IM tools for some aspects <strong>of</strong><br />
their <strong>work</strong>, given the ground that these tools has gained as personal communication infrastructure among<br />
the students. However, whereas most students in our study were users <strong>of</strong> a particular IM service,<br />
individual preferences vary. In the projects, we occasionally saw individual students refusing to use<br />
certain tools on principle (e.g. by being against proprietary s<strong>of</strong>tware) while the rest <strong>of</strong> the team wanted to<br />
use those tools. In such cases, the teams might benefit from their supervisor‟s mitigation to find good<br />
solutions.<br />
For the organizing <strong>of</strong> SE project courses, we see the following implications <strong>of</strong> students‟ competence with<br />
IM tools:<br />
<br />
<br />
<br />
<strong>The</strong> student teams should be allowed to choose the collaboration technology for their projects, within<br />
the constraints <strong>of</strong> availability, the team‟s competence <strong>and</strong> the customer‟s requirements.<br />
Teams should be required to present their supervisor with an account <strong>of</strong> their practices <strong>of</strong> using<br />
instant messaging, <strong>and</strong> the supervisor should use this information as a basis for providing process<br />
guidance.<br />
In projects heavily depending on distributed <strong>work</strong>, the use <strong>of</strong> instant messaging chat to aid such <strong>work</strong><br />
should be discussed with the team.<br />
Example 2 illustrated that informal collaboration between the team <strong>and</strong> the supervisor may be<br />
successfully conducted over IM. We believe that keeping IM open as a communication channel between<br />
team <strong>and</strong> supervisor can encourage students to get early in touch when they are in need <strong>of</strong> help, which<br />
makes it possible to solve problems before they escalate <strong>and</strong> before too much time is spent on being<br />
stuck. <strong>The</strong> viability <strong>of</strong> this solution depends on the supervisor‟s willingness to use IM in communication<br />
with the student <strong>and</strong> to be available for consultation on a fairly short, or at least explicitly agreed-upon,<br />
notice. We see the following general implication for the organizing <strong>of</strong> SE project courses:<br />
<br />
Course staff may use instant messaging as a collaboration tool for providing informal supervision or<br />
otherwise communicate informally with the project students. Expected availability (e.g. response<br />
time) should be explicitly addressed as the use <strong>of</strong> IM for collaboration is agreed with the team.<br />
Example 3 illustrates the general challenge <strong>of</strong> stakeholder communication <strong>and</strong>, in particular,<br />
communication with the customer, in project <strong>work</strong>. Successful stakeholder communication depends on<br />
project teams‟ ability to share the right information in the right format at the right time. Where there are<br />
language <strong>and</strong>/or cultural barriers, the challenges <strong>of</strong> communication increase. As Frode from Example 3<br />
put it in an interview after completion <strong>of</strong> their project: “We did not know if the customer was joking or if<br />
he was angry with us!”. <strong>The</strong> choice <strong>and</strong> use <strong>of</strong> collaboration technology strongly impacted on the<br />
communication. IM chat with the use <strong>of</strong> st<strong>and</strong>ard functionality (i.e. text only) provided opportunities for<br />
130<br />
22
immediate feedback in the heated discussion while conveying a sense <strong>of</strong> informality <strong>and</strong> providing<br />
limited opportunities for the communicating parties to monitor each other‟s reactions (e.g. through tone <strong>of</strong><br />
voice <strong>and</strong> body language). In Example 3, the mismatch between the communication needs <strong>and</strong> the<br />
affordances <strong>of</strong> the communication medium was particularly severe. <strong>The</strong> Molotov cocktail <strong>of</strong> problematic<br />
factors included the team <strong>and</strong> customer never having met, not knowing each other, having different native<br />
language <strong>and</strong> cultural background (including organizational/university culture – one being hierarchical<br />
<strong>and</strong> the other flat), having different conceptions <strong>of</strong> responsibilities connected to the customer role in the<br />
project, <strong>and</strong> the team being in a situation <strong>of</strong> severe problems (being far behind schedule while important<br />
resources were missing).<br />
<strong>The</strong> team may be unable to identify a potential mismatch between the choice <strong>of</strong> communication medium,<br />
the purpose <strong>of</strong> communication <strong>and</strong> the form <strong>of</strong> communicating. In some cases, as in Example 3, the<br />
customer might have suggested an inadequate arrangement for communication. A similar <strong>and</strong> typical<br />
example, related to another type <strong>of</strong> collaboration tool, is when the customer insists on the use <strong>of</strong> email for<br />
all student enquiries but turns out to be very slow in answering his email.<br />
<strong>The</strong> biggest <strong>and</strong> most general problem illustrated in Example 3, however, is that IM is being used for<br />
formal collaboration. Many project issues require a formal meeting, properly prepared <strong>and</strong> documented,<br />
or formal, thought-through, written communication, with copies <strong>and</strong> attachments. Instant messaging can<br />
be seen to encourage an informal tone <strong>and</strong> comments which are not meant to be revisited for purposes <strong>of</strong><br />
documentation, even if logging is possible. If IM is used in team-customer collaboration, the logs should<br />
be kept by the team for purposes <strong>of</strong> process documentation, <strong>and</strong> the customer should be informed about<br />
this usage. For instance, the IM log from which Example 3 is drawn was an important part <strong>of</strong> the<br />
documentation <strong>of</strong> the project process <strong>and</strong> used by course staff in the evaluation <strong>of</strong> the project. <strong>The</strong> IM log<br />
illustrated some reasons for the disappointing project result.<br />
In contrast to Example 3, our findings from the SE project course also included examples <strong>of</strong> successful<br />
use <strong>of</strong> IM in team-customer collaboration, enabling teams to have quick feedback from a remote customer<br />
(See Section 4.2) on mainly technical aspects <strong>of</strong> s<strong>of</strong>tware development. In this case, the purpose <strong>and</strong><br />
issues <strong>of</strong> IM conversation were limited <strong>and</strong> clear.<br />
Implications for the organizing <strong>of</strong> SE project courses <strong>of</strong> the challenges <strong>of</strong> using IM for customer<br />
communication include:<br />
<br />
<br />
<br />
<br />
Given the high project risk associated with unsuccessful customer communication, project<br />
supervisors should make sure that they are aware <strong>of</strong>, <strong>and</strong> accordingly require the team to account for,<br />
how collaboration with stakeholders is conducted in the team, including the use <strong>of</strong> collaboration<br />
technology, e.g. instant messaging tools. If the customer has asked for an arrangement <strong>of</strong><br />
communication highly likely to cause problems for the team, the supervisor might provide the team<br />
with advice on how to change the arrangement or point out what are the pitfalls.<br />
With respect to formal communication, in general, instant messaging should be used with care. <strong>The</strong><br />
participants should ensure that the needs <strong>of</strong> formal communication are met, for instance that there is<br />
enough information to provide the social awareness necessary for the parties to underst<strong>and</strong> each<br />
other‟s views <strong>and</strong> feelings (e.g. satisfaction, anger) when important issues are discussed. This might<br />
not be possible over instant messaging alone if the parties are not familiar with each other from<br />
previous face-to-face interaction. Also, the parties should make sure that the communication is<br />
documented in a way suitable for archiving <strong>and</strong> later retrieval.<br />
If any collaboration with the customer takes place over instant messaging, the conversations should<br />
be logged, in agreement with the customer<br />
If the customer is motivated to use instant messaging to be able to give rapid feedback on technical<br />
issues, the team should use the opportunity.<br />
Many <strong>of</strong> the implications outlined in this section address issues that apply to the use <strong>of</strong> collaboration<br />
technology in general to support <strong>work</strong> in the SE student teams, not only to the use <strong>of</strong> instant messaging.<br />
Even if most teams in our study seemed to be capable <strong>of</strong> choosing adequate media for various aspects <strong>of</strong><br />
cooperative <strong>work</strong>, we suggest that it be made a topic in the introductory presentation <strong>of</strong> the project course.<br />
131<br />
23
Also, initial sharing <strong>of</strong> experience among students in class may spread good ideas across teams <strong>and</strong><br />
further help course staff underst<strong>and</strong> students‟ current practices. In the project course <strong>of</strong> our study we have<br />
good experience with conducting a mid-term <strong>work</strong>shop in which students meet across teams to discuss<br />
various aspects <strong>of</strong> their project <strong>work</strong>. <strong>The</strong> more or less successful use <strong>of</strong> collaboration tools is an<br />
appropriate issue for such discussion. To facilitate fruitful exchange <strong>of</strong> experience, students may be asked<br />
to provide scenarios <strong>of</strong> collaboration from their projects, accounting for how issues were resolved <strong>and</strong><br />
how collaboration technology was used. To illustrate the actual collaboration practices, logs documenting<br />
the collaboration may be used. IM logs are an example <strong>of</strong> documentation that for reasons <strong>of</strong> privacy may<br />
be inadequate for publicly presenting a team‟s collaborative practices. However, for the teams internally,<br />
the logs may be a valuable source <strong>of</strong> information if they want to reflect on <strong>and</strong> learn from their own<br />
project process (Krogstie 2009).<br />
6 CONCLUSION<br />
In this paper, we have presented findings on the use <strong>of</strong> instant messaging tools in SE student projects.<br />
Used in the projects, the tools serve to support aspects <strong>of</strong> the <strong>work</strong> process <strong>and</strong> are means <strong>of</strong> providing<br />
realistic <strong>work</strong> experience as intended in project-based <strong>learning</strong>. Findings from our studies indicate that the<br />
teams‟ use <strong>of</strong> instant messaging for team-internal collaboration, whether for quick coordination or lengthy<br />
exchanges over ongoing, distributed <strong>work</strong>, is frequently successful.<br />
In organizing <strong>and</strong> supervising SE student projects we should be aware <strong>of</strong> students‟ existing competence in<br />
using IM tools <strong>and</strong> the type <strong>of</strong> collaboration typically happening with the aid <strong>of</strong> these tools. To gain the<br />
necessary underst<strong>and</strong>ing <strong>of</strong> the project process required for guidance <strong>of</strong> the specific team, we should ask<br />
the team to account for its collaboration practices, including the use <strong>of</strong> collaboration tools such as IM<br />
tools.<br />
<strong>The</strong> use <strong>of</strong> IM to support collaboration with project stakeholders is potentially challenging. Based on our<br />
findings, we propose that the student teams be advised to restrict the use <strong>of</strong> instant messaging to informal<br />
collaboration. Much <strong>of</strong> the customer collaboration in the SE projects may be considered formal <strong>and</strong><br />
requires the use <strong>of</strong> other types <strong>of</strong> collaboration technology.<br />
Finally, we have pointed to the use <strong>of</strong> collaboration tools such as IM in the SE projects as a topic worth<br />
discussion <strong>and</strong> <strong>reflection</strong>. On the level <strong>of</strong> the project course, this may be done in introductory lectures <strong>and</strong><br />
<strong>work</strong>shops. In addition, the use <strong>of</strong> IM <strong>and</strong> other collaboration tools should be discussed internally in each<br />
project team to help the teams gain insights about their own <strong>work</strong>.<br />
While we have limited the focus <strong>of</strong> this paper to student projects in the domain <strong>of</strong> s<strong>of</strong>tware engineering,<br />
our results have relevance for project-based <strong>learning</strong> in other domains in which instant messaging tools<br />
are used to aid the project <strong>work</strong>.<br />
A limitation to our study is the lack <strong>of</strong> data combining observation <strong>of</strong> IM use with access to the resulting<br />
logs. While IM logs in combination with post-hoc interviews with project participants provided rich data<br />
about the role <strong>of</strong> IM in supporting <strong>work</strong> processes, observation <strong>of</strong> the <strong>work</strong> taking place as the logs were<br />
produced would have provided valuable additional insights. However, we have been able to obtain IM<br />
logs reflecting collaboration processes identified by the teams as important to their projects. Another<br />
limitation to the study lies in the analysis <strong>of</strong> the interviews forming the basis for the findings on patterns<br />
<strong>of</strong> IM use in the student teams. While the material has been thoroughly examined, the analysis would<br />
have benefited from systematic categorization <strong>of</strong> the answers about instant messaging, organized e.g. in a<br />
table. For instance, in the paper, we combined the data from the 2006-2008 interviews based on the<br />
informal finding that there was no change in the patterns <strong>of</strong> IM use over these years. This finding <strong>and</strong><br />
approach could have been better substantiated with quantification <strong>of</strong> the data.<br />
As further <strong>work</strong> we plan to introduce the topic <strong>of</strong> collaboration tool use in <strong>reflection</strong> <strong>work</strong>shops in a SE<br />
project course to increase students‟ consciousness <strong>of</strong> this aspect <strong>of</strong> their <strong>work</strong> practice. We will evaluate<br />
the outcomes to see if this deliberate focus leads to insights in the student teams about how the different<br />
collaboration tools, including instant messaging, should be used to support their project <strong>work</strong>.<br />
132<br />
24
Acknowledgements<br />
Thanks to Monica Divitini <strong>and</strong> John Krogstie for feedback on ideas <strong>and</strong> drafts, <strong>and</strong> to the SE students for<br />
sharing their views <strong>and</strong> instant messaging logs.<br />
References<br />
Blumenfeld, P. C., E. Soloway, et al. (1991). "Motivating Project-Based Learning: Sustaining the Doing,<br />
Supporting the Learning." Educational Psychologist 26(3 & 4): 369-398.<br />
Bygstad, B., B. Krogstie, et al. (2006). Scaffolding Project Based Learning with the Rational Unified<br />
Process. Experience from 5 years <strong>of</strong> Student Projects in S<strong>of</strong>tware Engineering. NOKOBIT 2006,<br />
Molde, Norway, Tapir.<br />
Bygstad, B., B. R. Krogstie, et al. (2009). "Learning from achievement: scaffolding student projects in<br />
s<strong>of</strong>tware engineering " International Journal <strong>of</strong> Net<strong>work</strong>ed <strong>and</strong> Virtual Organizations 6(2).<br />
Carstensen, P. H. <strong>and</strong> K. Schmidt (2002 (1999)). “<strong>Computer</strong> Supported Cooperative Work: New<br />
Challenges to Systems Design”. H<strong>and</strong>book <strong>of</strong> Human Factors/Ergonomics. K. Itoh. Tokyo,<br />
Asakura Publishing.<br />
Castells, M. (2000). <strong>The</strong> Internet Galaxy: Reflections on the Internet, business, <strong>and</strong> Society. Oxford,<br />
Oxford University Press.<br />
Cherry, S. M. (2002). "IM means business." Spectrum, IEEE 39(11): 28 - 32<br />
Garrett, R. K. <strong>and</strong> J. N. Danziger (2008). "IM=Interruption Management? Instant Messaging <strong>and</strong><br />
Disruption in the Workplace." Journal <strong>of</strong> <strong>Computer</strong>-Mediated Communication 13(1).<br />
Grinter, R. E. <strong>and</strong> M. Eldridge (2003). Wan2tlk?: Everyday Text Messaging. CHI2003, Ft.Lauderdale,<br />
Florida, USA, ACM.<br />
Grinter, R. E. <strong>and</strong> L. Palen (2002). Instant Messaging in Teen Life. CSCW'02, New Orelans, Louisiana,<br />
USA, ACM.<br />
Gutwin, C., R. Penner, et al. (2004). Group Awareness in Distributed S<strong>of</strong>tware Development. CSCW'04,<br />
Chicago, Illinois, USA, ACM.<br />
Huang, A. H. <strong>and</strong> D. C. Yen (2003). "Usefulness <strong>of</strong> instant messaging among young users: Social vs.<br />
<strong>work</strong> perspective." Human Systems Management 22(2): 63-72.<br />
Isaacs, E., A. Walendowski, et al. (2002). <strong>The</strong> Character, Functions, <strong>and</strong> Styles <strong>of</strong> Instant Messaging in<br />
the Workplace. CSCW'02, New Orleans, Louisiana, USA, ACM.<br />
Klein, H. K. <strong>and</strong> M. M. Myers (1999). "A Set <strong>of</strong> Principles for Conducting <strong>and</strong> Evaluating Interpretive<br />
Field Studies in Information Systems." MIS Quarterly 23(1): 67-94.<br />
Krogstie, B. (2008). Power through brokering. OSS participation in SE projects. ICSE 2008, Leipzig,<br />
IEEE <strong>Computer</strong> Society.<br />
133<br />
25
Krogstie, B. <strong>and</strong> B. Bygstad (2007). Cross-Community Collaboration <strong>and</strong> Learning in Customer-Driven<br />
S<strong>of</strong>tware Engineering Student Projects. Twentieth Conference on S<strong>of</strong>tware Engineering<br />
Education <strong>and</strong> Training (CSEE&T), Dublin, IEEE <strong>Computer</strong> Society.<br />
Krogstie, B. R. (2008). <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. COOP, Carry-le-Rouet, Provence,<br />
France, Institut d‟Etudes Politiques d‟Aix-en-Provence.<br />
Krogstie, B. R. (2009). A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> utilizing historical<br />
data in collaborative tools. EC-TEL 2009, Nice, France, Springer.<br />
Krogstie, B. R. (2009). Using Project Wiki History to Reflect on the Project Process 42nd Hawaii<br />
International Conference on System Sciences, Big Isl<strong>and</strong>, Hawaii, IEEE <strong>Computer</strong> Society.<br />
Lancaster, S., D. C. Yen, et al. (2007). "<strong>The</strong> selection <strong>of</strong> instant messaging or e-mail. College students'<br />
perspective for computer communication." Information Management & <strong>Computer</strong> Security 15(1):<br />
5-22.<br />
Lewis, C. <strong>and</strong> B. Fabos (2005). "Instant messaging, literacies, <strong>and</strong> social identities." Reading Research<br />
Quarterly 40(4): 470-501.<br />
McMillan, W. W. (1999). What Leading Practitioners Say Should Be Emphasized in Students' S<strong>of</strong>tware<br />
Engineering Projects. Conference on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training, New<br />
Orleans, Louisiana, USA, IEEE <strong>Computer</strong> Society.<br />
Nardi, B. A., S. Whittaker, et al. (2000). Interaction <strong>and</strong> Outeraction: Instant Messaging in Action.<br />
CSCW'00, Philadelphia, PA, USA, ACM.<br />
Stephens, K. K. (2008). "Optimizing Costs in Workplace Instant Messaging." IEEE Transactions on<br />
Pr<strong>of</strong>essional Communication 51(4): 369-380.<br />
Thomas, J. (2000). A review <strong>of</strong> research on project-based <strong>learning</strong>. Novato, CA, <strong>The</strong> Buck Institute for<br />
Education.<br />
Yin, R. K. (2003). Case Study Research. Design <strong>and</strong> Methods. Third Edition, SAGE Publications.<br />
134<br />
26
Research paper P4<br />
Title:<br />
<strong>The</strong> wiki as an integrative tool in project <strong>work</strong><br />
Author:<br />
Birgit Krogstie<br />
Published in:<br />
Proceedings <strong>of</strong> COOP 2008<br />
Pages:<br />
12<br />
Complete reference:<br />
Krogstie, B. R. (2008). <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. COOP, Carryle-Rouet,<br />
Provence, France, 20-23 May. Institut d’Etudes Politiques d’Aix-en-<br />
Provence.<br />
135
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
136
<strong>The</strong> wiki as an integrative tool in project <strong>work</strong><br />
Birgit R. Krogstie<br />
1<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology,<br />
Sem Sæl<strong>and</strong>s vei 5-7, NO-7491 Trondheim , Norway<br />
birgitkr @idi.ntnu.no<br />
Abstract. <strong>The</strong> paper provides insights on how wikis support project <strong>work</strong> <strong>and</strong><br />
what characteristics <strong>of</strong> wikis make them adequate for this purpose. <strong>The</strong> findings<br />
are based on a case study <strong>of</strong> s<strong>of</strong>tware engineering projects in an educational<br />
setting. Project wikis are found to serve an integrative role along several<br />
dimensions <strong>of</strong> project <strong>work</strong>, enabled by the flexibility <strong>and</strong> automatic support for<br />
capturing history <strong>of</strong>fered by the technology. <strong>The</strong> findings demonstrate that a<br />
project wiki can serve as a knowledge repository, a means for staging the<br />
project, a coordination mechanism, <strong>and</strong> a shared <strong>work</strong>space. To many projects<br />
in need <strong>of</strong> project management <strong>and</strong> collaboration support, project wikis should<br />
be seen as an attractive, lightweight, all-purpose alternative.<br />
Keywords: wikis, cooperation technology, project <strong>work</strong>, s<strong>of</strong>tware engineering<br />
1 Introduction<br />
”When we had a wiki up <strong>and</strong> running in the first place, we wanted to use it for as<br />
much as possible” (Project manager, Team A)<br />
Wikis are lightweight technology supporting easy creation <strong>and</strong> maintenance <strong>of</strong> web<br />
pages. <strong>The</strong> flexible structuring <strong>of</strong> contents is a particular feature making wikis<br />
adequate <strong>and</strong> popular as knowledge repositories. A wiki may be used for keeping the<br />
shared information resources <strong>of</strong> a group such as an organizational unit or <strong>and</strong> internet<br />
community, allowing the members add <strong>and</strong> modify contents. <strong>The</strong> contents <strong>of</strong> a wiki<br />
may be shared with the world or restricted to a group or community [1].<br />
Lightweight project management (PM) solutions are <strong>of</strong>ten sought when ‘heavy’<br />
PM is not feasible, as seen in open source s<strong>of</strong>tware development. Louridas [2] points<br />
to potential advantages <strong>of</strong> wikis to support s<strong>of</strong>tware engineering (SE) <strong>work</strong>:<br />
comprehensive versioning <strong>and</strong> change control, the provision <strong>of</strong> a shared <strong>work</strong>space<br />
for producing <strong>and</strong> storing project documentation, <strong>and</strong> the possibility <strong>of</strong> having input<br />
from many participants in a more effective way than e.g. over email. Louridas argues<br />
that wikis are especially useful as discussion media, the advantages over discussion<br />
lists being general user-friendliness <strong>and</strong> the fact that discussion items can be placed in<br />
context. Further, he points to the democratic aspect: a project wiki gives a voice to<br />
everybody in the project. Chao [3] presents survey findings <strong>of</strong> innovative forms <strong>of</strong><br />
wiki use in student SE teams.<br />
137
As course staff in SE project courses we have seen many student teams using wikis<br />
during the last couple <strong>of</strong> years. <strong>The</strong> project wikis contain or provide links to various<br />
project artifacts, convey team-internal as well as information for external stakeholders<br />
<strong>and</strong> generally seem to reflect a mixture <strong>of</strong> administrative, development-related <strong>and</strong><br />
socio-emotional aspects <strong>of</strong> project <strong>work</strong>. Existing research literature does not address<br />
in depth how project wikis aid different aspects <strong>of</strong> SE practice. For instance, the <strong>work</strong><br />
<strong>of</strong> Chao on wiki use in SE projects identifies different types <strong>of</strong> use, but only on an<br />
overview level based. Louridas addresses the potential <strong>of</strong> wikis in SE projects without<br />
presenting empirical findings on actual use.<br />
<strong>The</strong> objective <strong>of</strong> the <strong>work</strong> presented in this paper has been to gain insights on the<br />
actual <strong>work</strong> practices involving wiki use in SE teams <strong>and</strong> the purposes for which the<br />
wikis are used in these teams. In what follows, we provide an account <strong>of</strong> related<br />
research. Next, we describe our case <strong>and</strong> research method. We present findings on<br />
how the wikis in our case serve to integrate aspects <strong>of</strong> project <strong>work</strong> along several<br />
dimensions, providing illustrative examples. Finally, we relate this to a set <strong>of</strong> different<br />
roles played by the wikis in the projects, before concluding the paper.<br />
2 Related <strong>work</strong><br />
Whereas our focus is on the role <strong>of</strong> wikis in SE <strong>work</strong>, it is worthwile to examine<br />
the body <strong>of</strong> research on wikis in educational contexts. Much <strong>of</strong> this research turns on<br />
the use <strong>of</strong> one wiki for an entire course [4] <strong>and</strong>/or possible pedagogical extensions <strong>of</strong><br />
wiki functionality addressing teacher-student interaction, e.g. [5, 6]. We believe that<br />
there are important differences between course wikis <strong>and</strong> project wikis (while some<br />
wikis may be hybrids). <strong>The</strong> analytic distinction provided by Riel <strong>and</strong> Polin [7]<br />
between practice-based, knowledge-based <strong>and</strong> task-based <strong>learning</strong> communities<br />
captures a major point: A SE student project is mainly task-based. In the type <strong>of</strong><br />
project considered in our case, which is industry projects with a high degree <strong>of</strong><br />
independence on part <strong>of</strong> the project teams, the project wikis are established by the<br />
team to support their <strong>work</strong> towards project completion <strong>and</strong> the cooperative<br />
development <strong>of</strong> a product. <strong>The</strong> wiki is typically not intended to be re-used across<br />
projects (as might be the case if knowledge building or sustained practice within a<br />
course or organization were the objectives). In other words, to underst<strong>and</strong> the usage <strong>of</strong><br />
a SE project wiki with particular heed to the perspective <strong>of</strong> its users, it is reasonable to<br />
view it as collaboration technology used by a project team even if the particular<br />
setting is that <strong>of</strong> a university course.<br />
Brown et al. [8] addressed how wikis can be used as a collaboration tool among<br />
students <strong>of</strong> ethnography. <strong>The</strong> type <strong>of</strong> wiki in question resembles a course wiki more<br />
than a project wiki, but there are aspects <strong>of</strong> its use which resembles project wiki usage<br />
in SE teams. In particular, Brown et al studied the use <strong>of</strong> wikis for collaboration over<br />
the writing <strong>of</strong> field notes, students sharing, comparing <strong>and</strong> discussing their notes in<br />
the wiki. Findings address e.g. the willingness <strong>of</strong> wiki users to share personal<br />
fieldnotes on an open wiki <strong>and</strong> the very limited editing <strong>of</strong> what is perceived as others’<br />
documents in the wiki. A finding <strong>of</strong> particular relevance to SE <strong>work</strong> is the combined<br />
distributed <strong>and</strong> co-located wiki use found in the groups <strong>of</strong> ethnography students. This<br />
138
type <strong>of</strong> use can be seen as an example <strong>of</strong> how teams in a situation <strong>of</strong> partly<br />
distributed, nomadic <strong>work</strong> [9] find ways to utilize the flexibility <strong>of</strong> the wiki to serve<br />
their situated needs.<br />
In the context <strong>of</strong> project <strong>work</strong> in general, the usefulness <strong>of</strong> wikis has increasingly<br />
been recognized [2, 10]. Singh et al. [11] show how interdisciplinary projects can use<br />
social bookmarking to support personalized access to wiki content. As mentioned in<br />
the introduction, Louridas sees a number <strong>of</strong> advantages <strong>of</strong> wikis in SE <strong>work</strong> [2].<br />
Wikis can be seen as a form <strong>of</strong> collaboration technology fitting the philosophy <strong>and</strong><br />
practices <strong>of</strong> free <strong>and</strong> open source development [12]. Mullick <strong>and</strong> colleagues [10]<br />
introduced a wiki in a simulated global s<strong>of</strong>tware development project to improve<br />
coordination among distributed teams, finding that a common portal to project<br />
information was necessary to coordinate the design <strong>work</strong> across teams.<br />
Chao’s study <strong>of</strong> wiki use in student SE teams [3] unveiled satisfaction with the use<br />
<strong>of</strong> wikis as a lightweight collaboration tool. In what follows, we exp<strong>and</strong> on Chao’s<br />
<strong>work</strong> to find what it is about project wikis that serves SE teams so well.<br />
3 Case <strong>and</strong> research method<br />
We conducted a field study during the spring 2007 among s<strong>of</strong>tware engineering<br />
student teams developing s<strong>of</strong>tware solutions for external customers. Teams <strong>of</strong> 3-5<br />
students <strong>work</strong> half time on their projects throughout their sixth (<strong>and</strong> last) semester,<br />
drawing on <strong>and</strong> integrating knowledge from other courses in the s<strong>of</strong>tware engineering<br />
program. <strong>The</strong> overarching <strong>learning</strong> objective is to gain experience with development<br />
<strong>work</strong> in a team for a customer, the management <strong>of</strong> the team-customer relationship<br />
intended to be a major challenge <strong>and</strong> source <strong>of</strong> <strong>learning</strong> for the students. Each team<br />
has an external customer providing a development task unique to that team. <strong>The</strong><br />
customer is responsible for providing necessary information on requirements <strong>and</strong><br />
feedback on the development <strong>of</strong> the product. Further, each team has a supervisor from<br />
course staff providing guidance on the project process. Each team receives one grade,<br />
which is set in cooperation between the supervisor <strong>and</strong> an external examiner.<br />
Status reporting <strong>and</strong> bi-weekly meetings with the supervisor are m<strong>and</strong>atory, as are<br />
the delivery <strong>of</strong> two preliminary versions <strong>of</strong> a project report before the final project<br />
delivery. Apart from this, strong independence on part <strong>of</strong> the students is encouraged.<br />
It is up to the teams to organize themselves, to manage the communication <strong>and</strong><br />
relationship with project stakeholders, <strong>and</strong> to choose <strong>and</strong> adapt an appropriate<br />
development methodology / process model as well as collaboration <strong>and</strong> development<br />
tools. <strong>The</strong> choices made by the teams depend on the customer’s preferences or<br />
requirements, team members’ prior knowledge <strong>and</strong> experience, team member’s wish<br />
to learn, <strong>and</strong> the availability <strong>of</strong> the technology (e.g. through university license, as<br />
freeware, or provided by the customer).<br />
Four out <strong>of</strong> the nine teams in the 2007 cohort made use <strong>of</strong> project wikis, <strong>and</strong> three<br />
teams (hereby denoted team A, team T <strong>and</strong> team D) were active wiki users throughout<br />
the entire project. <strong>The</strong> project deliverables included a report (electronic <strong>and</strong> on paper)<br />
<strong>and</strong> source code, but the wiki was not itself a deliverable.<br />
139
Following an interpretive approach [13], we collected data on these teams through<br />
participant observation, interviews with the various stakeholders (teams, supervisors<br />
<strong>and</strong> customers), examination <strong>of</strong> project documentation <strong>and</strong> project wikis. <strong>The</strong> data has<br />
been coded based on themes emerging through the analysis, without explicit reference<br />
to one theoretical frame<strong>work</strong>, but with the use <strong>of</strong> core CSCW concepts as deemed<br />
useful. Students from the teams A, T <strong>and</strong> D were encouraged to provide feedback on<br />
the full version <strong>of</strong> the paper, but we did not succeed in having their opinion.<br />
4 Findings <strong>and</strong> analysis<br />
In our analysis, we investigate wiki usage in teams A, T <strong>and</strong> D, assuming the<br />
projects to be similar enough to make it possible to draw a picture <strong>of</strong> wiki usage<br />
across cases. All the teams had the necessary technical competence to solve their task.<br />
<strong>The</strong>y were successful with the project in terms <strong>of</strong> achieving a good grade <strong>and</strong><br />
delivering a product to their customer in accordance with specifications. <strong>The</strong>y had a<br />
project manager being a project blog sponsor/enthusiast, creating the blog <strong>and</strong> taking<br />
the main responsibility for updating it (while all other team members also contributed,<br />
to varying degrees). <strong>The</strong>y used their project wiki as a shared project <strong>work</strong>space <strong>and</strong><br />
main channel for updated <strong>and</strong> <strong>of</strong>ficial project status information, as will be elaborated<br />
<strong>The</strong>y regularly updated a to-do-list <strong>and</strong> other memo/task-oriented lists on the wiki.<br />
<strong>The</strong>y password-protected the wiki or avoided informing widely about the web<br />
address, but made the wiki accessible to customer <strong>and</strong> supervisor. Finally, they used a<br />
version management system separate from the wiki for the s<strong>of</strong>tware development<br />
Some more detail on each <strong>of</strong> the teams illustrates the differences <strong>and</strong> provides<br />
context for some <strong>of</strong> the findings presented in this section:<br />
Team D appeared very structured <strong>and</strong> ambitious. <strong>The</strong>re were 3 members, one <strong>of</strong><br />
which was absent during a large part <strong>of</strong> the project period due to serious illness. <strong>The</strong><br />
two others were the lead developers <strong>and</strong> relatively old students, experienced with<br />
project management <strong>and</strong> s<strong>of</strong>tware engineering. <strong>The</strong> project <strong>work</strong> was almost solely<br />
done in a distributed fashion. <strong>The</strong> customer was located in the same city as the<br />
students <strong>and</strong> the university. <strong>The</strong> development task was technically challenging. <strong>The</strong><br />
application used for generating the project wiki was DokuWiki. After a while during<br />
the project period the team included a wiki plug-in to manage project activities.<br />
Team A had 5 members who were mostly collocated in their project <strong>work</strong>. <strong>The</strong>ir<br />
customer was located in a different part <strong>of</strong> the country during the first part <strong>of</strong> the<br />
project. <strong>The</strong> application used for the project wiki was MediaWiki. After the mid-term<br />
project report delivery, the project manager developed a script for collaboratively<br />
writing the final project report in the wiki <strong>and</strong> having the report auto-generated.<br />
Team T had 5 members. <strong>The</strong>ir project <strong>work</strong> was largely collocated. <strong>The</strong> customer<br />
was located in the same city as the team <strong>and</strong> the university. T. Among the three teams<br />
in our study, team T had the greatest number <strong>of</strong> personal <strong>and</strong> emotional expressions in<br />
their wiki, which was a MediaWiki. <strong>The</strong> team used only st<strong>and</strong>ard wiki functionality.<br />
In the three project teams, we found that the project wikis serve a role in the teams’<br />
efforts to address <strong>and</strong> balance considerations typically competing for project<br />
resources <strong>and</strong> attention. <strong>The</strong> overarching perspective <strong>of</strong> our analysis thus evolved to<br />
140
ecome integration, <strong>and</strong> we identified several dimensions along which this integration<br />
seemed to receive wiki support. <strong>The</strong> dimensions are, as will be seen, interwoven.<br />
4.1 Wiki usage: Integrating across the social <strong>and</strong> the task-oriented<br />
In this section we elaborate on the use <strong>of</strong> project wikis to convey task-oriented as<br />
well as social/emotional information. We address how to-do-lists are used, how<br />
individual project contributions are made visible, <strong>and</strong> how project status information<br />
is conveyed through the structuring <strong>of</strong> the main page.<br />
<strong>The</strong> wiki main page is used for reflecting project status, in particular remaining<br />
tasks, but also the social <strong>and</strong> emotional state <strong>of</strong> affairs.<br />
“<strong>The</strong> wiki quickly proved itself to be an invaluable tool for information management<br />
<strong>and</strong> collaboration, with the front page being used primarily as an editable bulletin<br />
board describing the current messages/tasks, <strong>and</strong> sub pages holding the actual<br />
content.” [..] (interview, team D)<br />
<strong>The</strong> teams were very selective in terms <strong>of</strong> what should go into the main page.<br />
Issues that could be resolved elsewhere, e.g. in weekly meetings or over msn, need<br />
not, <strong>and</strong> should not – according to the students - be put in the wiki. To-do-lists <strong>and</strong><br />
news bulletins on the main page were for conveying <strong>and</strong> reading the project status.<br />
Each team kept a to-do-list on their main page. In addition, there were various<br />
other lists, their number, position <strong>and</strong> purpose varying across teams <strong>and</strong> between<br />
phases <strong>of</strong> the project. Examples include individual to-do-lists, a project news bulletin,<br />
<strong>and</strong> separate to-do-lists for the project report <strong>and</strong> the testing.<br />
“What is put on sub pages tends to be forgotten. What is current, what is to be done<br />
now, should be on the main page. What is most recent should be on top <strong>of</strong> the list.”<br />
(interview, team A)<br />
<strong>The</strong> lists <strong>of</strong> remaining tasks continuously changed throughout the projects, growing<br />
<strong>and</strong> shrinking depending on the phase <strong>of</strong> the project <strong>and</strong> the distance to the next<br />
deadline. At need, a dedicated list (e.g. for the final report) was branched out.<br />
<strong>The</strong> process <strong>of</strong> assigning tasks to team members was typically done in face-to-face<br />
<strong>work</strong> settings, but also through the wiki: from the list <strong>of</strong> currently unresolved tasks, a<br />
team member could just pick one <strong>and</strong> <strong>work</strong> on it, in case <strong>of</strong> one <strong>of</strong> the teams by<br />
moving the item to his/her individual list in the wiki. To-do-lists were thus not only<br />
supplementing (on a day-to-day level) but in fact partly substituting other project<br />
planning or development artifacts such as a ticket system for managing s<strong>of</strong>tware<br />
changes. Team T saw the wiki as “an easy way out” in terms <strong>of</strong> change management:<br />
“<strong>The</strong> ticket system does have some features that might be better; you can.. integrate it<br />
with SVN, then, the versioning control, when you make a new change, you may write<br />
something particular that makes the ticket system pick it up, <strong>and</strong> see that things have<br />
been changed, but it becomes so advanced, that ..people do not bother to get into it,<br />
you know. And then it is easier only to have a wiki-page.” (interview, team T)<br />
Team D, on the other h<strong>and</strong>, integrated a wiki plug-in for the management <strong>of</strong><br />
activity lists, thus taking a more structured approach to s<strong>of</strong>tware changes in their wiki.<br />
Strikethroughs were extensively used as a means for making visible the completion<br />
<strong>of</strong> tasks in the wiki to-do-lists. An item with a strikethrough typically remained in the<br />
list until the overall task (e.g. a preliminary report) had been completed.<br />
141
Project logos <strong>and</strong> other visual material were used to achieve a certain<br />
personalization <strong>of</strong> the web page, related to a team identity, but also to individuals. An<br />
example <strong>of</strong> the latter is that the wiki responsible team member <strong>of</strong> team T included a<br />
link to a website which provided a new r<strong>and</strong>om picture <strong>of</strong> a cat with each click – cats<br />
being a strong personal area <strong>of</strong> interest to the team member.<br />
Emotional expressions in the wiki relating to the management <strong>of</strong> tasks ranged from<br />
individual team members’ comments by the descriptions <strong>of</strong> tasks (“Finally!”, “What<br />
to do with the graphics?”) via the project manager’s use <strong>of</strong> red, large-size fonts for<br />
urgent messages to pictures representing the general feeling <strong>of</strong> a deadline approaching<br />
(Fig.1):<br />
Fig. 1: Part <strong>of</strong> the wiki main page <strong>of</strong> team T towards final deadline.<br />
Strong outbursts were <strong>of</strong>ten quickly diminished or removed by their creators in<br />
subsequent wiki versions. <strong>The</strong> flexibility to easily remove material thus possibly<br />
encourages the (sometimes brief) exposure <strong>of</strong> such material on the wiki.<br />
Individual to-do-lists indicated that individuals had, or had taken, responsibility for<br />
specific tasks. Initials by tasks in a common task list served the same purpose. Effort<br />
made was indicated by strikethroughs <strong>and</strong> comments, e.g. on difficulties with the task.<br />
A cake list was launched on team T’s main page as a kind <strong>of</strong> pillory. Members<br />
being late for meetings had their names put in the list. <strong>The</strong> team thus made visible that<br />
it was not acceptable to break team rules, <strong>and</strong> the rule breakers received a small<br />
penalty by having their name exposed. At the same time, they were <strong>of</strong>fered the<br />
opportunity to make amends through a social contribution (making a cake for the<br />
team). Viewed as part <strong>of</strong> the account <strong>of</strong> the status <strong>of</strong> the project, names on the cake<br />
list indicated (minor) project management challenges, <strong>and</strong> their being addressed.<br />
A wiki main page can easily be re-structured while other pages remain unchanged.<br />
<strong>The</strong> main pages in our study were resized along with the phases <strong>and</strong> needs <strong>of</strong> the<br />
142
project. Typically, a clean-up was performed after a completed deadline, <strong>and</strong> spin-<strong>of</strong>fs<br />
(e.g. new sub-pages or separate headings on the main page) resulted from increased<br />
activity in certain areas or the recognition <strong>of</strong> new needs <strong>and</strong> objectives.<strong>The</strong><br />
restructuring thus reflects a coordination <strong>of</strong> tasks, but it also has an emotional side,<br />
e.g. making visible that something has been accomplished, or that the team is turning<br />
over a new leaf. A fully packed main page, maybe slightly more chaotic than usual,<br />
expresses something different about how the project is currently experienced by its<br />
members than a sparsely populated page with short to-do-lists.<br />
4.2 Wiki usage: Integrating team-internal <strong>and</strong> team-external information<br />
From the above account <strong>of</strong> task-oriented <strong>and</strong> social/emotional use <strong>of</strong> the wiki, it is<br />
evident that the project wikis were used for team-internal purposes. <strong>The</strong> wikis were<br />
however also used for providing external stakeholders (customer, supervisor <strong>and</strong><br />
possibly others) with information about the status <strong>of</strong> the project (e.g. project plan,<br />
status report, latest revision <strong>of</strong> requirements specification).<br />
Making the wiki externally available means, to some extent, providing a desirable<br />
image <strong>of</strong> the project. A nicely structured wiki with tasks gradually getting their<br />
strikethroughs convey an image <strong>of</strong> a process under control. <strong>The</strong> emotional<br />
expressions found on the pages, however, reveal emotional sides <strong>of</strong> the project <strong>work</strong><br />
to external stakeholders. <strong>The</strong> teams in our study found this acceptable, but allowed for<br />
different amounts <strong>of</strong> informal <strong>and</strong> emotionally laden contents on their wikis.<br />
Team D saw the wiki as very central in communicating project status to the<br />
customer. In their final report, the wiki was placed in the middle <strong>of</strong> a figure showing<br />
the collaboration structure <strong>of</strong> the project. <strong>The</strong> wiki was shown as the link between the<br />
team <strong>and</strong> the other stakeholders. To the surprise <strong>of</strong> team D, however, the customer, in<br />
his evaluation report on the project, expressed satisfaction with everything in the<br />
project except that he would have wanted more face to face meetings instead <strong>of</strong><br />
updates only through the wiki. In the case <strong>of</strong> teams A <strong>and</strong> T, the supervisor told the<br />
team that he wanted them to use a wiki <strong>and</strong> give him access to it, which most likely<br />
influenced the teams’ choice to use wiki technology.<br />
When asked if they would have liked to share wiki contents with other teams in the<br />
course, the teams in our cases were generally positive. <strong>The</strong> project manager <strong>of</strong> team A<br />
however thought that they might not want to display task lists full <strong>of</strong> uncompleted<br />
tasks because it would reveal a lack <strong>of</strong> control. This comment reflects the general<br />
competitive attitude among the project teams in the course.<br />
Communication through the wiki seemed to happen largely one-way, from the<br />
team to other stakeholders. Customer <strong>and</strong> supervisor generally provided feedback<br />
through other channels. In the case <strong>of</strong> team A, however, the customer was situated in<br />
another city during the first half <strong>of</strong> the project <strong>and</strong> frequently visited the wiki. If there<br />
was something <strong>of</strong> particular importance in the wiki, the team would send him email<br />
about it. <strong>The</strong> team had a couple <strong>of</strong> conferences with the customer over msn <strong>and</strong><br />
Skype, <strong>and</strong>, the customer had been visiting the wiki before the meetings, examining<br />
for instance the current version <strong>of</strong> the requirement specification. Further, the customer<br />
maintained a separate page in the project wiki containing non-functional requirements<br />
for the system under development. In the case <strong>of</strong> team T, one <strong>of</strong> two customer<br />
143
contacts once made comments directly in the requirements specification in the project<br />
wiki. He later explained that he found it interesting to try out new technology for this<br />
purpose. A typical customer attitude to project wiki pages, however, seems to be that<br />
they are the students’ area, even when other stakeholders are given write permission.<br />
4.3 Wiki usage: Integrating across artifacts<br />
Major project artifacts, e.g. minutes, status reports, plans <strong>and</strong> implementation<br />
related information, were typically accessible from the main page. Also, the wikis<br />
contained links to relevant web pages with data e.g. about the customer or a tool.<br />
<strong>The</strong> project task <strong>of</strong> the teams in our study was a SE task, <strong>and</strong> coding was done in<br />
separate environments. Wiki integration with code versioning was very limited. None<br />
<strong>of</strong> the teams were familiar with wikis providing support for such integration.<br />
Lightweight <strong>and</strong> more manual solutions were however helpful. Team A did not use a<br />
separate versioning system, but used their customer’s code repository, to which access<br />
was restricted. <strong>The</strong> project manager uploaded modified source code to wiki pages for<br />
each member to download locally (as the team could not consecutively run their code<br />
on a shared server), using msn to notify when a new download should be made.<br />
Text formatting in wikis is not compatible with commonly used text editors. <strong>The</strong><br />
teams accordingly ‘wikified’ documentation by adding the necessary tags to<br />
unformatted text. This was time-consuming, annoying <strong>and</strong> infeasible for the full<br />
project report. None <strong>of</strong> the teams were familiar with wiki functionality for generating<br />
printable pages from wiki pages. <strong>The</strong> low-threshold alternative to integrating the<br />
report into the wiki was to make it accessible through a link. One team developed a<br />
script converting report chapters in the wiki into Latex, making it possible to perform<br />
joint editing within the wiki <strong>and</strong> generate a printable version at any time.<br />
<strong>The</strong> wikis are used in combination with other collaboration tools to <strong>work</strong> on <strong>and</strong><br />
discuss project artifacts. <strong>The</strong> teams are very clear about which tools are used for<br />
which purpose. Discussions typically happen face-to-face or over instant messaging.<br />
Chat among distributed programmers is done in parallel with programming <strong>work</strong>, <strong>and</strong><br />
unresolved <strong>work</strong> issues might be added to the wiki to-do-list during the <strong>work</strong> session.<br />
Email is preferred for formal communication, particularly when the customer is<br />
involved <strong>and</strong> there is a need for documentation.<br />
5 Discussion<br />
One way <strong>of</strong> accounting for the richness <strong>and</strong> versatility <strong>of</strong> the observed integrative<br />
wiki usage in the SE projects is to distinguish between different roles served by the<br />
wikis in the projects. We find that the project wikis simultaneously serve as<br />
knowledge repositories, stages, coordination mechanisms, <strong>and</strong> shared <strong>work</strong>spaces. In<br />
the light <strong>of</strong> these roles, we will elaborate on current wiki usage.<br />
144
5.1 <strong>The</strong> wiki as a knowledge repository<br />
Our findings indicate that the project wikis serve a role as knowledge repositories<br />
for the teams <strong>and</strong> other stakeholders. In all the wikis <strong>of</strong> our study, project artifacts <strong>and</strong><br />
useful resources were included or linked in, “so we know where to find it”. We may<br />
regard this as ‘traditional’ usage <strong>of</strong> wiki technology.<br />
Apart from the current version <strong>of</strong> the wiki at any point, historical information is<br />
stored in the wiki. Each wiki version can be seen as forming a point, <strong>of</strong> greater or<br />
lesser significance, along a project trajectory. Trajectory has been described as “(1)<br />
the course <strong>of</strong> any experienced phenomenon as it evolves over time [..] <strong>and</strong> (2) the<br />
actions <strong>and</strong> interactions contributing to its evolution” [14](p.53-54). <strong>The</strong> development<br />
over time <strong>of</strong> certain aspects <strong>of</strong> the project, or artifacts such as the s<strong>of</strong>tware product,<br />
may be seen as forming sub-trajectories that are, to some degree, represented in the<br />
wiki. Ideas developing into artifacts can be seen as trajectories <strong>of</strong> interaction,<br />
reflecting social as well as technical/task-oriented aspects <strong>of</strong> the project process.<br />
5.2 <strong>The</strong> wiki as a stage<br />
<strong>The</strong> teams provide certain types <strong>of</strong> information on the wiki while withholding other<br />
information. This can be seen as providing a certain image <strong>of</strong> the project, internally<br />
<strong>and</strong> externally. This is particularly evident from the main page, being loaded with<br />
status information <strong>of</strong> a technical as well as social <strong>and</strong> emotional nature. <strong>The</strong> latter<br />
type <strong>of</strong> material becomes somewhat restricted through the teams’ practice <strong>of</strong> not<br />
conducting discussions in the wikis (in spite <strong>of</strong> the potential benefits [2]), <strong>and</strong> through<br />
the fairly quick moderation or removal <strong>of</strong> some material.<br />
Internally in the team, the ‘staged’ picture <strong>of</strong> the project not only provides a status<br />
update but also potentially serves a role in conveying a project identity for team<br />
members to share. <strong>The</strong> inclusion <strong>of</strong> project logos <strong>and</strong> informal material such as<br />
humorous pictures may contribute to a feeling <strong>of</strong> project identity.<br />
For external use, a pr<strong>of</strong>essional appearance might be <strong>of</strong> greater importance.<br />
Information on the wiki constitutes part <strong>of</strong> the controlled information flow between<br />
project stakeholders. This is important e.g. in managing customer expectations. <strong>The</strong><br />
team’s version <strong>of</strong> project status as shown on the wiki should be regarded as one-way<br />
information in need <strong>of</strong> some channel for feedback (whether in the wiki or in another<br />
channel), as indicated by customer D’s wish for more face-to-face interaction.<br />
In the projects <strong>of</strong> our study, the project wikis were not deliverables, whether to<br />
customer or course staff. <strong>The</strong> wikis were however accessible to the supervisors, who<br />
played an important role in the formal evaluation <strong>of</strong> the team. Our data are not<br />
sufficient for making inferences about how this particular aspect <strong>of</strong> the course design<br />
affected wiki usage, e.g. the allowance for informal <strong>and</strong> unstructured contents.<br />
5.3 <strong>The</strong> wiki as a coordination mechanism<br />
We have seen how the to-do-lists provide a means for planning, monitoring <strong>and</strong> replanning<br />
project <strong>work</strong> in a very flexible way. <strong>The</strong> linkability <strong>and</strong> malleability <strong>of</strong> the<br />
145
tool [15] may be seen as major enablers <strong>of</strong> this usage, in combination with the history<br />
function which makes it possible to capture a picture <strong>of</strong> the process over time as part<br />
<strong>of</strong> the current status.<br />
<strong>The</strong> to-do-lists have a team-internal coordinative role, but also play a role in<br />
coordination with external stakeholders who monitor project status in order to act<br />
upon it. We argue that is not only the strictly task-oriented aspects but also the social<br />
<strong>and</strong> emotional side <strong>of</strong> the lists that constitute <strong>and</strong> inform articulation <strong>work</strong> in the<br />
teams, providing cues about how to appropriately approach current challenges.<br />
Louridas suggested that wikis are good for project democracy. In our study, we see<br />
that the project manager, being the major wiki sponsor in the team, uses the wiki to<br />
set or visualize the agenda for the project, thus manifesting his power. All team<br />
members are wiki users, to varying degrees, making visible their contributions. We<br />
may only speculate that the wiki <strong>of</strong>fers better visibility for the viewpoints <strong>of</strong> those<br />
who fail to get their voice through in face-to-face meetings.<br />
5.4 <strong>The</strong> wiki as shared <strong>work</strong>space<br />
<strong>The</strong> wiki constitutes a <strong>work</strong>space providing shared access to various artifacts. <strong>The</strong><br />
inherent wiki functionality for h<strong>and</strong>ling access conflicts to pages or page sections is<br />
appreciated by the teams in our study. However, we have seen that the current use <strong>of</strong><br />
wikis for collaborative writing is so far limited. Taken from team members’<br />
viewpoints on how they would want to use their wikis, it seems that lack <strong>of</strong><br />
knowledge <strong>of</strong>, or access to, adequate technology (e.g. for converting between formats)<br />
was a major reason why e.g. the project reports were not written in the wikis directly.<br />
Some <strong>work</strong>space awareness [16] is achieved as the wiki is used in combination<br />
with other development <strong>and</strong> collaboration tools, e.g. instant messaging. Also, the<br />
history <strong>of</strong> the wiki <strong>and</strong> information on the main page on who recently did what,<br />
provides awareness <strong>of</strong> team members’ <strong>work</strong>. <strong>The</strong> project wikis may be seen as<br />
socially translucent [17], providing social cues about team members <strong>and</strong> their effort –<br />
both what they actually did <strong>and</strong>, to some extent, how they felt about it. A wiki main<br />
page with its to-do-list, individual comments <strong>and</strong> strikethroughs provides visibility<br />
<strong>and</strong> accountability <strong>of</strong> team members as well as <strong>of</strong> the team as a whole.<br />
Turning to the project trajectory in context <strong>of</strong> the shared <strong>work</strong>space, the concept <strong>of</strong><br />
locale [18] is useful. Fitzpatrick made interaction trajectory one <strong>of</strong> the core aspects <strong>of</strong><br />
the locale – “the place constituted in the ongoing relationship between people in a<br />
particular social world <strong>and</strong> the ‘site <strong>and</strong> means’ they se to meet their interactional<br />
needs, i.e. the space together with the resources <strong>and</strong> things available there” (p.9) A<br />
project wiki may be regarded as part <strong>of</strong> these ‘site <strong>and</strong> means’ in the project, <strong>and</strong> the<br />
history function <strong>of</strong> the wiki as a potentially valuable resource for reviewing the<br />
project trajectory in order to learn from it <strong>and</strong>/or <strong>work</strong> on its future course.<br />
A final remark should be made on the emotional <strong>and</strong> social contents <strong>of</strong> the wikis in<br />
our study. <strong>The</strong>se contents are mostly about reflecting <strong>and</strong> supporting a task-oriented<br />
project process by providing some social awareness. In the teams, social interaction<br />
generally takes place on other arenas. Accordingly, the project wikis should not be<br />
considered as ‘social s<strong>of</strong>tware’.<br />
146
6. Conclusion<br />
We have shown how project teams make use <strong>of</strong> wikis in a way integrating aspects<br />
<strong>of</strong> their project <strong>work</strong>. <strong>The</strong> wikis are simultaneously knowledge repositories, means <strong>of</strong><br />
staging the projects, coordination mechanisms <strong>and</strong> shared <strong>work</strong>spaces.<br />
We believe that a core issue in determining the appropriateness <strong>of</strong> using a project<br />
wiki in a specific project is how much predefined structure is sought for the<br />
management <strong>of</strong> the project. This relates to the choice <strong>of</strong> development methodology<br />
<strong>and</strong> life<strong>cycle</strong> model. Agile development <strong>and</strong> open source projects tend to favour<br />
lightweight <strong>and</strong> flexible tools [12]. Further, project size, heterogeneity <strong>and</strong> degree <strong>of</strong><br />
distributed vs. collocated <strong>work</strong> impacts on the overall complexity <strong>and</strong> need for<br />
computerized coordination mechanisms [19]. <strong>The</strong> development <strong>of</strong> safety-critical<br />
projects require a more rigid development process than projects involving less risk.<br />
We suggest that the use <strong>of</strong> a project wiki as a lightweight project management tool<br />
is appropriate for projects that are not safety critical. We see project wikis as relevant<br />
for SE student projects, small SE projects (particularly if <strong>work</strong> is at least partially<br />
distributed), teams within larger SE projects, <strong>and</strong> open source SE projects.<br />
Our findings demonstrated the potential <strong>of</strong> a wiki to serve many roles in SE project<br />
<strong>work</strong> besides that <strong>of</strong> coordination mechanism. If the wiki were to be supplemented<br />
with a project management tool in which coordination was h<strong>and</strong>led, it is difficult to<br />
predict how the wiki might serves its other roles (knowledge repository, stage, shared<br />
<strong>work</strong>space). This topic might be pursued through empirical research.<br />
Whereas our main research agenda has been to provide a descriptive account <strong>of</strong><br />
current wiki usage, we would like to point to alternative usage <strong>and</strong> further research.<br />
First, there is a potential for utilizing the knowledge repository <strong>of</strong> the individual<br />
projects across projects. Students express interest in looking at earlier project reports<br />
to learn from them, <strong>and</strong> this is relevant to industry settings as well. If a project is<br />
staged to be accessible <strong>and</strong> intelligible to all stakeholders, it is likely to be <strong>of</strong> potential<br />
value to other project teams. Project wikis may generally have a place in the broader<br />
picture <strong>of</strong> knowledge management in an organization. Second, given the effort<br />
currently made by the teams to create <strong>and</strong> maintain informative wiki pages, there is a<br />
potential to conduct project-external communication through this channel to a larger<br />
extent than what we see in the teams today. Generally, the practice <strong>of</strong> using the<br />
customer’s preferred channel (typically email) to convey wiki contents by including<br />
links to the relevant wiki pages, seems to be successful as a way <strong>of</strong> providing a slight<br />
information push. Third, better opportunities for collaborative writing were requested<br />
by our teams. Having the wiki tool manage versioning <strong>and</strong> access conflicts <strong>and</strong> <strong>of</strong>fer<br />
functionality to generate nicely formatted reports would significantly increase the<br />
value <strong>of</strong> wikis as shared <strong>work</strong>space. Some existing wikis provide such solutions.<br />
Taken from our findings, there is only limited use <strong>of</strong> historical information in the<br />
SE project wikis. History stored in the wiki can be utilized for purposes <strong>of</strong> <strong>reflection</strong>,<br />
e.g. in post-mortem project reviews. Our research continues along this line.<br />
147
Acknowledgements<br />
This <strong>work</strong> was funded through the MOTUS2 project at NTNU. Monica Divitini <strong>and</strong><br />
John Krogstie provided valuable comments on draft versions <strong>of</strong> the paper.<br />
References<br />
1. Leuf, B. <strong>and</strong> W. Cunningham, <strong>The</strong> Wiki Way - quick collaboration on the Web. 2001:<br />
Addison-Wesley.<br />
2. Louridas, P., Using Wikis in S<strong>of</strong>tware Development. IEEE S<strong>of</strong>tware, 2006.<br />
3. Chao, J. Student Project Collaboration using Wikis. in 20th Conference on S<strong>of</strong>tware<br />
Engineering & Training (CSEET'07). 2007. Dublin, Irel<strong>and</strong>.<br />
4. Xu, L., Project the wiki way: Using wiki for computer science course project management.<br />
Journal <strong>of</strong> Computing Sciences in Colleges 2007. 22:(6).<br />
5. Lund, A. <strong>and</strong> O. Smørdal. Is <strong>The</strong>re a Space for the Teacher in a WIKI? in Proceedings <strong>of</strong><br />
the 2006 international symposium on Wikis WikiSym '06 2006.<br />
6. Reinhold, S. WikiTrails: augmenting Wiki structure for collaborative, interdisciplinary<br />
<strong>learning</strong>. in International Symposium On Wikis 2006.<br />
7. Riel, M. <strong>and</strong> L. Polin, Online Learning Communities. Common Ground <strong>and</strong> Critical<br />
Differences in Designing Technical Environments, in Designing for Virtual Communities in<br />
the Service <strong>of</strong> Learning, S.A.K. Barab, Rob; Gray, James H., Editor. 2004, Cambridge<br />
University Press: Cambridge. p. 16-50.<br />
8. Brown, B., et al. Seeing Ethnographically: Teaching ethnography as part <strong>of</strong> CSCW. in<br />
ECSW'07. 2007. Limerick, Irel<strong>and</strong>.<br />
9. Bødker, S., et al. Technology for Boundaries. in GROUP'03. 2003. Sanibel Isl<strong>and</strong>, Florida,<br />
USA: ACM<br />
10. Mullick, N., et al. Siemens Global Studio Project: Experiences Adopting an Integrated GSD<br />
Infrastructure. in International Conference on Global S<strong>of</strong>tware Engineering. 2007. Munich,<br />
Germany: IEEE Press.<br />
11. Singh, A.V., A. Wombacher, <strong>and</strong> K. Aberer. Personalized Information Access in a Wiki<br />
Using Structured Tagging. in OTM. 2007: Springer.<br />
12. Gutwin, C., R. Penner, <strong>and</strong> K. Schneider. Group Awareness in Distributed S<strong>of</strong>tware<br />
Development. in CSCW. 2004. Chicago, Illinois, USA: ACM.<br />
13. Klein, H.K. <strong>and</strong> M.M. Myers, A Set <strong>of</strong> Principles for Conducting <strong>and</strong> Evaluating<br />
Interpretive Field Studies in Information Systems. MIS Quarterly, 1999. 23:(1) p. 67-94.<br />
14. Strauss, A., Continual permutations <strong>of</strong> action. 1993, New York: Aldine de Gruyter.<br />
15. Schmidt, K. <strong>and</strong> C. Simone, Coordination Mechanisms: Towards a Conceptual Foundation<br />
<strong>of</strong> CSCW Systems Design. <strong>Computer</strong> Supported Cooperative Work, 1996. 5 p. 155-200.<br />
16. Gutwin, C. <strong>and</strong> S. Greenberg, A Descriptive Frame<strong>work</strong> <strong>of</strong> Workspace Awareness for Real-<br />
Time Groupware. <strong>Computer</strong> Supported Cooperative Work, 2002. 11:(3-4) p. 411-446.<br />
17. Erickson, T. <strong>and</strong> W.A. Kellogg, Social Translucence: An Approach to Designing Systems<br />
that Support Social Processes. ACM TOCHI, 2000. 7:(1) p. 59-83.<br />
18. Fitzpatrick, G., <strong>The</strong> Locales Frame<strong>work</strong>: Underst<strong>and</strong>ing <strong>and</strong> Designing for Wicked<br />
Problems. <strong>The</strong> Kluwer International series on <strong>Computer</strong> Supported Cooperative Work, ed.<br />
R. Harper. 2003: Kluwer Academic Publishers.<br />
19. Trac home page, http://trac.edgewall.org/. Last accessed 17.April 2008.<br />
20. Carstensen, P.H. <strong>and</strong> K. Schmidt, <strong>Computer</strong> Supported Cooperative Work: New Challenges<br />
to Systems Design, in H<strong>and</strong>book <strong>of</strong> Human Factors/Ergonomics, K. Itoh, Editor. 2002<br />
(1999), Asakura Publishing: Tokyo.<br />
148
Research paper P5<br />
Title:<br />
Using Project Wiki History to Reflect on the Project Process<br />
Author:<br />
Birgit Krogstie<br />
Published in:<br />
Proceedings <strong>of</strong> HICSS 2009<br />
Pages:<br />
10<br />
Complete reference:<br />
Krogstie, B. R. (2009). Using Project Wiki History to Reflect on the Project Process<br />
42nd Hawaii International Conference on System Sciences (HICSS’42), Big Isl<strong>and</strong>,<br />
Hawaii, 5-8 Jan. IEEE <strong>Computer</strong> Society.<br />
149
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
150
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
Using Project Wiki History to Reflect on the Project Process<br />
Birgit R. Krogstie<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
birgitkr@idi.ntnu.no<br />
Abstract<br />
In this paper, we address the potential <strong>of</strong> project<br />
wikis to support <strong>reflection</strong> on the project process<br />
through participants’ reconstruction <strong>of</strong> the project<br />
trajectory. Drawing on a case study <strong>of</strong> s<strong>of</strong>tware<br />
engineering project <strong>work</strong>, we illustrate how<br />
information on project history can be found in a<br />
project wiki <strong>and</strong> may be used to support recall <strong>of</strong> <strong>and</strong><br />
<strong>reflection</strong> on the process. We discuss implications for<br />
postmortem project reviews by considering how the<br />
utilization <strong>of</strong> project wikis could addresses some<br />
challenges to successful reviews. We also propose<br />
extended wiki functionality that can be used to make a<br />
more selective review <strong>of</strong> project history based on usertagged<br />
contents.<br />
1. Introduction<br />
S<strong>of</strong>tware engineering (SE) <strong>work</strong> is acknowledged<br />
to be complex, as elaborated in the literature [1, 2] on<br />
computer-supported cooperative <strong>work</strong> (CSCW) <strong>and</strong><br />
demonstrated by the frequent SE failures [3].<br />
Underst<strong>and</strong>ing the history <strong>of</strong> a SE project is essential to<br />
SE practice in two ways: In day-to-day coordination<br />
<strong>and</strong> project management, <strong>and</strong> in <strong>learning</strong> from the<br />
project experience. Such <strong>learning</strong> can be seen both as<br />
an intrinsic aspect <strong>of</strong> <strong>work</strong> [4, 5] <strong>and</strong> as means for<br />
achieving organizational <strong>learning</strong> within <strong>and</strong> across<br />
projects [6-8]. We might also add that in the field <strong>of</strong><br />
education, project <strong>work</strong> is acknowledged to be a<br />
particularly effective setting for <strong>learning</strong> [9].<br />
This paper addresses how project wiki history can<br />
support postmortem <strong>reflection</strong> on the project process.<br />
<strong>The</strong> paper does not address why a team should choose<br />
to use a project wiki, or how project wikis compare to<br />
other project management tools. We take as a starting<br />
point the acknowledgement that many teams currently<br />
do use project wikis for lightweight project<br />
management. We argue that there is more to these<br />
wikis than what can be seen in day-to-day project<br />
management: there is a lot <strong>of</strong> historical information<br />
with a potential to shed light on the project history.<br />
In proposing this approach to improving SE<br />
postmortem reviews, we address some recognized<br />
challenges to making such reviews successful. We also<br />
present the design <strong>of</strong> a wiki extension which can be<br />
used to support the approach. A prototype <strong>of</strong> the wiki<br />
extension has been developed <strong>and</strong> is briefly presented<br />
in the paper along for a scenario <strong>of</strong> its use. We do not<br />
provide complete guidelines for a postmortem review<br />
utilizing wiki history.<br />
<strong>The</strong>oretically, we lean on the <strong>work</strong> <strong>of</strong> A. Strauss<br />
<strong>and</strong> the concept <strong>of</strong> trajectory. We see an SE teams<br />
reconstruction <strong>of</strong> its project trajectory as essential to<br />
<strong>reflection</strong> on the project process.<br />
<strong>The</strong> paper is structured in the following way: <strong>The</strong><br />
Background section addresses some core concepts <strong>and</strong><br />
related <strong>work</strong>. In Section 3 we present our case <strong>and</strong><br />
research method. In Section 4 we draw on a case study<br />
to show that project wikis have a potential to support<br />
<strong>reflection</strong> on the project process. In Section 5, we<br />
discuss how project wikis can be used to address some<br />
challenges to postmortem reviews. A tool to improve<br />
the usefulness <strong>of</strong> project wikis in such reviews is<br />
described in Section 6. Section 7 addresses limitations<br />
<strong>and</strong> further <strong>work</strong> <strong>and</strong> concludes the paper.<br />
2. Background<br />
Three issues provide a background for the<br />
following chapters <strong>and</strong> are addressed in this section:<br />
How trajectories play a role in <strong>reflection</strong>, the practice<br />
<strong>and</strong> challenges <strong>of</strong> postmortem reviews in SE, <strong>and</strong> the<br />
use <strong>of</strong> wikis as lightweight project management tools<br />
in SE projects.<br />
2.1. <strong>The</strong> role <strong>of</strong> trajectories in <strong>reflection</strong><br />
Dewey defined reflective thought as 'active,<br />
persistent, <strong>and</strong> careful consideration <strong>of</strong> any belief or<br />
supposed form <strong>of</strong> knowledge in the light <strong>of</strong> the grounds<br />
that support it <strong>and</strong> the further conclusions in the<br />
direction <strong>of</strong> which it tends' [10, p.133]. In this paper,<br />
we consider <strong>reflection</strong> as a process <strong>of</strong> assessing<br />
something (e.g. a product or a sequence <strong>of</strong> events) in<br />
978-0-7695-3450-3/09 $25.00 © 2009 IEEE<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
151<br />
1
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
respect to some quality <strong>and</strong> with respect to alternatives.<br />
For instance, in <strong>reflection</strong> on a process, the congruence<br />
with, or deviation from, some other process, whether<br />
real (e.g. the process <strong>of</strong> a previous project) or imagined<br />
(e.g. an alternative course <strong>of</strong> the same process, or the<br />
ideal process as prescribed in a model) will typically<br />
be considered.<br />
We subscribe to the view that human activity is<br />
basically social [10, 11] <strong>and</strong> that reality is socially<br />
constructed [12]. Individual members <strong>of</strong> a social world,<br />
such as a project team, belong to different social<br />
worlds <strong>and</strong> have varying degrees <strong>of</strong> involvement in<br />
these worlds over time [11, 13]. Reflecting on project<br />
experience involves the alignment <strong>of</strong> different<br />
participants perspectives.<br />
Strauss [11] refers to Mead on the issue <strong>of</strong> how a<br />
social world reflects on its history: <strong>The</strong> temporal<br />
spans <strong>of</strong> group life mean that the aims <strong>and</strong> aspirations<br />
<strong>of</strong> group endeavor are subject to reviewal <strong>and</strong><br />
recasting. Likewise past activities come to be viewed<br />
in new lights, through reappraisal <strong>and</strong> selective<br />
recollection. ..History, whether that <strong>of</strong> a single person<br />
or <strong>of</strong> a group, signifies a coming back at self (Mead<br />
1936). Strauss uses the term trajectory to denote<br />
the course <strong>of</strong> any experienced phenomenon as it<br />
evolves over time. Events as they unfold, e.g. in a<br />
project, form a trajectory, <strong>and</strong> from each point in time<br />
there are many possible trajectories onwards.<br />
Taking the experienced phenomenon into<br />
consideration, different individuals trajectories will be<br />
different even when they largely comprise the same<br />
events. What actually happened, is a question <strong>of</strong><br />
interpretation <strong>and</strong> reconstruction. Also, trajectory is<br />
given meaning in a particular situation <strong>of</strong> use [14]. For<br />
instance, the same sequence <strong>of</strong> information elements<br />
presented after an intermediary deadline in a project<br />
will be seen in a different way if used for purposes <strong>of</strong><br />
coordination rather than to learn from the process. In<br />
sum, there is a lot <strong>of</strong> trajectory context which is not<br />
available from its representation (e.g. a textual<br />
description). Strictly, then, a trajectory is not a<br />
sequence <strong>of</strong> single elements <strong>of</strong> information as<br />
(re)presented, but as used. For practical purposes in<br />
what follows, however, we will frequently refer to<br />
trajectory as the trajectory as represented, e.g. a set <strong>of</strong><br />
information elements, chronologically ordered. Doing<br />
so, we keep in mind that the representation will gain<br />
meaning through situated sense-making the essence<br />
<strong>of</strong> the sought-after <strong>reflection</strong>.<br />
2.2. A case <strong>of</strong> <strong>reflection</strong>: SE postmortem<br />
reviews<br />
We now turn to <strong>reflection</strong> as part <strong>of</strong> s<strong>of</strong>tware<br />
engineering (SE) <strong>work</strong> practice. SE <strong>work</strong> is typically<br />
performed in teams <strong>and</strong> structured as projects. <strong>The</strong><br />
trajectory <strong>of</strong> a SE project is typically guided by a<br />
life<strong>cycle</strong> model incorporating phases <strong>and</strong> iterations.<br />
Options range from strictly <strong>and</strong> moderately structured<br />
models (e.g. traditional waterfall, RUP) to agile<br />
approaches (e.g. XP, SCRUM) [15]. A process model<br />
typically specifies or recommends a division <strong>of</strong> labor<br />
<strong>and</strong> a set <strong>of</strong> artifacts to be used <strong>and</strong> produced.<br />
Generally, deadlines for project deliverables heavily<br />
impact on the day-to-day experience <strong>of</strong> SE project<br />
<strong>work</strong>.<br />
To cope with the complexity <strong>of</strong> <strong>work</strong> <strong>and</strong> avoid the<br />
all too frequent failures in SE projects, efforts are<br />
needed to underst<strong>and</strong> project history beyond the status<br />
information needed for day-to-day articulation <strong>work</strong>.<br />
We will point to three major activities for which it is<br />
useful to review project history for purposes <strong>of</strong><br />
<strong>reflection</strong>. First, capturing <strong>and</strong> making explicit the<br />
design rationale has been seen as important to<br />
developers underst<strong>and</strong>ing <strong>of</strong> the system <strong>and</strong> as a<br />
source <strong>of</strong> <strong>learning</strong> for new team members [16, 17].<br />
Second, postmortem reviews are conducted to learn<br />
from experience within <strong>and</strong> across projects in an<br />
organization [18]. Third, supervision <strong>of</strong> SE project<br />
teams in educational settings involves gaining<br />
underst<strong>and</strong>ing <strong>of</strong>, <strong>and</strong> providing advice for, the project<br />
process [19].<br />
In what follows, we will focus on postmortem<br />
reviews. What is commonly known as a postmortem<br />
review is a collective <strong>learning</strong> activity organized for a<br />
project that has either finished a major activity or phase<br />
or has ended. Postmortems are conducted to reflect on<br />
what happened in the project in order to improve<br />
practice for individual participants <strong>and</strong> for the<br />
organization as a whole. A post mortem report is part<br />
<strong>of</strong> the review result. Several approaches to<br />
postmortems have been proposed in the literature <strong>and</strong><br />
taken into use in industry. <strong>The</strong> process typically takes<br />
from half a day to three days <strong>and</strong> is guided by a<br />
facilitator. Different approaches diverge over issues<br />
like home<strong>work</strong> or not, structured vs. open discussion,<br />
presence <strong>of</strong> management, appropriate output (e.g. how<br />
to make a useful report) <strong>and</strong> <strong>learning</strong> focus (tacit or<br />
explicit knowledge). <strong>The</strong> approaches provide lists <strong>of</strong><br />
steps in the review process [18]. A problem with<br />
postmortem reviews is that they are not heavily used in<br />
industry despite the recognized need to learn from<br />
s<strong>of</strong>tware engineering failures [20]. IT (Information<br />
Technology) specialists point to a number <strong>of</strong> barriers<br />
relating to set-up, data collection, data analysis <strong>and</strong><br />
knowledge sharing. Some <strong>of</strong> these challenges relate to<br />
the availability <strong>of</strong> relevant information on the project<br />
process <strong>and</strong> the means available for structuring that<br />
information. Three <strong>of</strong> the barriers [20] (pp.77-78) are:<br />
2<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
152
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
1) Insufficient integration with existing <strong>learning</strong><br />
systems: .[..] (e.g. reporting systems, quality control<br />
systems, <strong>and</strong> informal knowledge sharing).<br />
2) Simple information <strong>and</strong> complex issues:<br />
Information used in post mortem evaluations is limited<br />
<strong>and</strong> simplified compared to the complexity <strong>of</strong> the<br />
issues involved in system development.<br />
3) Information overload: Post mortem evaluations<br />
are overwhelming because there is too much to reflect<br />
on.<br />
2.3. <strong>The</strong> use <strong>of</strong> project wikis as lightweight<br />
project management tools<br />
To address current usage <strong>of</strong> project wikis <strong>and</strong> the<br />
types <strong>of</strong> information contained in the wikis, we refer to<br />
a case study reported in [21]. <strong>The</strong> case is an<br />
undergraduate project course in s<strong>of</strong>tware engineering<br />
(SE). <strong>The</strong> projects are organized to be as close to<br />
industry SE projects as possible. Teams <strong>of</strong> 3-5<br />
students <strong>work</strong> half time on their projects throughout<br />
their sixth (<strong>and</strong> last) semester in a Bachelor <strong>of</strong><br />
Informatics study program. Each team has an external<br />
customer providing a development task unique to that<br />
team. <strong>The</strong> customer is responsible for providing<br />
necessary information on requirements <strong>and</strong> feedback<br />
on the development <strong>of</strong> the product. Further, each team<br />
has a supervisor from course staff providing guidance<br />
on the project process. <strong>The</strong> project deliverables<br />
includes a report in three versions, as well as a <strong>work</strong>ing<br />
s<strong>of</strong>tware solution for the customer. Strong<br />
independence on part <strong>of</strong> the students is required. <strong>The</strong><br />
teams are responsible for managing stakeholder<br />
communication <strong>and</strong> project organization as well as<br />
choosing an appropriate process model <strong>and</strong><br />
development <strong>and</strong> collaboration tools. Four out <strong>of</strong> the<br />
nine teams in the 2007 cohort <strong>of</strong> our case made use <strong>of</strong><br />
project wikis, <strong>and</strong> three teams were active wiki users<br />
throughout the entire project. Two different wiki tools<br />
were used in the 2007 teams: DokuWiki <strong>and</strong><br />
MediaWiki.<br />
An exploratory case study was made by this author<br />
on the role <strong>of</strong> the wikis in these projects [21]. <strong>The</strong><br />
researcher was course coordinator, but did not evaluate<br />
the projects. Access to the wikis was granted for<br />
purposes <strong>of</strong> research. Data for the study also comprised<br />
interviews with the teams as well as various project<br />
documentation. In the study, the project wikis were<br />
found to serve an integrative role along several<br />
dimensions <strong>of</strong> project <strong>work</strong>, enabled by the flexibility<br />
<strong>and</strong> automatic support for capturing history <strong>of</strong>fered by<br />
the technology [21]. <strong>The</strong>se findings demonstrated that<br />
the project wikis simultaneously served as knowledge<br />
repositories, means for staging the projects,<br />
coordination mechanisms, <strong>and</strong> shared <strong>work</strong>spaces.<br />
<strong>The</strong> more purposes for which a project wiki has<br />
been used, the greater the potential to find useful<br />
information there about the corresponding aspects <strong>of</strong><br />
project <strong>work</strong>. For instance, if day-to-day coordination<br />
<strong>and</strong> allocation <strong>of</strong> tasks to team members is done with<br />
the aid <strong>of</strong> to-do-lists on the wiki main page, it is<br />
possible to use the wiki to reconstruct who did the<br />
<strong>work</strong> in various areas over time. If emotional<br />
expressions relating to the experience <strong>of</strong> project <strong>work</strong><br />
have been added to the wiki, the wiki may be used to<br />
reconstruct at least part <strong>of</strong> the story <strong>of</strong> what it was like<br />
to <strong>work</strong> in the project over time.<br />
A content page from a project wiki is shown in<br />
Figure 1. This is the upper part <strong>of</strong> the main page <strong>of</strong> the<br />
wiki referred to in Section 4, <strong>and</strong> the particular page<br />
revision is from April 16. 2008 at 12:13. A project<br />
main page typically contains overview information<br />
about the project, e.g. news bulletins <strong>and</strong> to-do-lists,<br />
<strong>and</strong> links to various information useful to the project<br />
team.<br />
Below the header there are links to previous <strong>and</strong><br />
next versions <strong>of</strong> the page, making it possible to<br />
chronologically traverse page versions. Surrounding<br />
the user-created contents <strong>of</strong> the page are menus<br />
providing access to various functionality, e.g to switch<br />
to editing mode, contribute to a discussion <strong>of</strong> the page,<br />
or view its history. Whereas this type <strong>of</strong> functionality<br />
is representative <strong>of</strong> wikis in general, the exact design<br />
shown is specific to DokuWiki.<br />
Figure 1. Wiki content page (Main page <strong>of</strong> team A)<br />
On basis <strong>of</strong> contents pages, the links between them,<br />
<strong>and</strong> the various page revisions, wiki tools <strong>of</strong>fer metapages<br />
with synthesized information. This includes<br />
overviews <strong>of</strong> all pages, the history <strong>of</strong> changes to one<br />
3<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
153
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
page, all revisions <strong>of</strong> a page, <strong>and</strong> the difference<br />
between two specified page revisions. <strong>The</strong> upper part<br />
<strong>of</strong> a history page is shown in Figure 2. <strong>The</strong> page<br />
revisions are listed chronologically, with links to each<br />
<strong>of</strong> them. From the list, it is possible to see who edited<br />
the page, when it was done, <strong>and</strong> what part <strong>of</strong> it was<br />
edited (if specified). Also, it is possible to see the<br />
difference, e.g. changes, between any two versions <strong>of</strong><br />
the page.<br />
Figure 2. History page <strong>of</strong> the main page in Figure 1<br />
For purposes <strong>of</strong> day-to-day project management,<br />
for which timely status information is essential,<br />
functionality for extracting synthesized information<br />
from the project wiki is useful. For instance, all the<br />
teams in our 2007 study reported that the history<br />
function in the wiki was occasionally used during the<br />
project to identify team members contribution.<br />
When the purpose <strong>of</strong> extracting information is to<br />
have a review <strong>of</strong> project history <strong>and</strong> reflect on the<br />
project process, the numerous content page revisions<br />
<strong>and</strong> the history pages providing overviews <strong>of</strong> these<br />
revisions are an interesting resource. In what follows,<br />
we will investigate the potential to use this resource for<br />
such a purpose.<br />
3. Case <strong>and</strong> research method<br />
Two <strong>of</strong> the post-mortem interviews conducted in<br />
the case study referred to in Section 2.3 were<br />
facilitated by the use <strong>of</strong> the teams project wiki. <strong>The</strong>se<br />
interviews turned out to shed light on the wiki as a<br />
means to support recall <strong>and</strong> <strong>reflection</strong> on the process<br />
after project completion. This issue was singled out as<br />
a separate research topic, <strong>and</strong> the resulting research is<br />
described in this paper.<br />
To present <strong>and</strong> substantiate our findings on wikifacilitated<br />
<strong>reflection</strong>, we have chosen to focus on one<br />
<strong>of</strong> the wiki-aided interviews. By focusing on a single<br />
case, we aim to provide a coherent account <strong>of</strong> how the<br />
traversal <strong>of</strong> wiki contents can shed light on a particular<br />
project process. <strong>The</strong> interview was conducted with the<br />
team manager <strong>of</strong> Project A one month after project<br />
completion. An interview guide with a number <strong>of</strong><br />
questions was used as a checklist in the interview, but<br />
issues were generally explored as they emerged in the<br />
conversation. <strong>The</strong> 96 versions <strong>of</strong> the wiki main page<br />
were chronologically traversed during the interview,<br />
starting with the first revision dated February 1. 2007,<br />
<strong>and</strong> ending with May 23. 2007, the day <strong>of</strong> project<br />
delivery. <strong>The</strong> project wiki was a DokuWiki [22].<br />
As the main page was the central page in the wiki<br />
in terms <strong>of</strong> project status information <strong>and</strong> coordination,<br />
the revisions <strong>of</strong> the main page became the basis for the<br />
interview. Links to other pages were followed<br />
whenever suggested by the interviewee <strong>and</strong> in some<br />
cases the interviewer. <strong>The</strong> interview lasted for 2 hours<br />
<strong>and</strong> was partially summarized in detail, partially<br />
transcribed. Wiki screenshots were added to the<br />
summary/transcript to make a coherent basis for<br />
analysis. Interview excerpts considered relevant for<br />
publication were translated from Norwegian to English<br />
at need. <strong>The</strong> wiki contents were originally in English.<br />
Data sources used in the analysis include the<br />
researchers thorough examination <strong>of</strong> the project wiki,<br />
various other project documentation from the team, a<br />
post-mortem interview with the team supervisor, the<br />
customers project evaluation, <strong>and</strong> an interview<br />
conducted with the entire team earlier in the semester.<br />
During analysis <strong>of</strong> the interview, we identified excerpts<br />
reflecting the interviewees recall <strong>of</strong>, or <strong>reflection</strong> on,<br />
the project trajectory with reference to wiki contents.<br />
<strong>The</strong> findings reported in the next section are based on<br />
these excerpts <strong>and</strong> the referred to wiki pages.<br />
4. Recall <strong>and</strong> <strong>reflection</strong> through project<br />
wiki traversal<br />
In the interview with the Project A manager<br />
(hereafter called TB), the interviewer went<br />
chronologically through revisions <strong>of</strong> the wiki main<br />
page, visiting other pages (content <strong>and</strong> history) at need.<br />
Interviewer <strong>and</strong> interviewee were sitting next to each<br />
other in front <strong>of</strong> the interviewers computer screen.<br />
In this section, we provide examples from the<br />
interview illustrating how information from the wiki is<br />
actively used in a process <strong>of</strong> recalling <strong>and</strong> reflecting on<br />
the project process. In each example, one or more<br />
4<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
154
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
content pages, history pages or a combination <strong>of</strong> both<br />
are used.<br />
<strong>The</strong> following contextual information may aid<br />
comprehension <strong>of</strong> the first example: <strong>The</strong> customer was<br />
located in another city throughout the project. As the<br />
team was to develop a module in a larger system, it<br />
was important that the team follow the customers<br />
coding conventions.<br />
4.1. Example 1: <strong>The</strong> introduction <strong>and</strong> use <strong>of</strong><br />
the customers page<br />
On the wiki main page, under the headline News<br />
bulletin, there is a bullet point with a link stating:<br />
Useful information <strong>and</strong> tidbits! Step right up <strong>and</strong> get it<br />
while its hot, folks. As TB clicks on the link, he<br />
enters the customers page. On top <strong>of</strong> the customers<br />
page is a contents list with the header Information<br />
from . <strong>The</strong> contents mainly address<br />
technical issues such as coding conventions. For<br />
instance, the first section is called If, foreach etc. <strong>and</strong><br />
starts with Always use curly braces when writing<br />
blocks. A code example is given in the section.<br />
TB goes to the history page. Looking at the<br />
information there, it can be seen that the page was<br />
created by TB in February, <strong>and</strong> that it has been updated<br />
by the customer only, in the period until early March.<br />
TB explains that he created the first version <strong>of</strong> the page<br />
based on an email from the customer with some<br />
guidelines for development. <strong>The</strong>n he had provided the<br />
customer with access rights to the wiki <strong>and</strong> a link to the<br />
customer page. After this, the customer continued<br />
using the page to provide the team with information,<br />
mostly non-functional system requirements <strong>and</strong><br />
guidelines about how to use a frame<strong>work</strong> written by<br />
the customer. <strong>The</strong> interviewer asks why there is no<br />
more information in the later phases <strong>of</strong> the project, <strong>and</strong><br />
TB explains that the team knew how things were to be<br />
done then.<br />
In this example, TB uses the content <strong>and</strong> history <strong>of</strong><br />
the customer page to account for a form <strong>of</strong><br />
collaboration with the (remote) customer <strong>and</strong> how it<br />
was successfully initiated by the team. <strong>The</strong><br />
interviewers follow-up question leads to <strong>reflection</strong> on<br />
the status <strong>of</strong> the project halfway in the project period:<br />
in TBs opinion, the team at that point was in control<br />
<strong>of</strong> non-functional requirements <strong>and</strong> mastered the way<br />
<strong>of</strong> doing development required by the customer.<br />
4.2. Example 2: Conveying urgency<br />
TB looks at the main page on May 3. at 22:36. <strong>The</strong><br />
first item in the News bulletin starts with the words<br />
UN-FREAKING-BELIEVABLY URGENT in upper<br />
case, red, boldface letters immediately drawing<br />
attention to that part <strong>of</strong> the page. In the next page<br />
revision, created 23:31, the font size is even larger.<br />
When the interviewer comments on this, TB says that<br />
the change shows the development <strong>and</strong> a few<br />
psychological reactions, sort <strong>of</strong>.<br />
In this example, changes between page revisions<br />
are a source <strong>of</strong> recall <strong>and</strong> <strong>reflection</strong>. Changes made<br />
from one hour to the next conveys a shapshot <strong>of</strong> the<br />
project; a taste <strong>of</strong> the experience <strong>of</strong> doing project <strong>work</strong><br />
under conditions <strong>of</strong> perceived urgency. In the example,<br />
TB points out how the web site reflects the feelings<br />
involved his feelings at the time, as it were.<br />
4.3. Example 3: A developing to-do-list<br />
TB looks at the main page as it looked on May 3. In<br />
the news bulletin, there is a new item on top with a link<br />
to a Final Report to-do list. TB explains that they had<br />
to branch out a task list. He goes to the Final Report<br />
TODO page <strong>and</strong> then to its history page. He notes that<br />
all revisions have been created by another team<br />
member <strong>and</strong> comments that this team member had<br />
main responsibility for updating the page. <strong>The</strong><br />
interviewer traverses many revisions <strong>of</strong> the page from<br />
the period May 3.-8. In these versions, new to-do-list<br />
items are steadily added <strong>and</strong> at one point the list is<br />
restructured to contain sub tasks. <strong>The</strong> tasks gradually<br />
become marked with strikethroughs. TB comments<br />
that some structural changes can be seen from the<br />
development <strong>of</strong> the list. He comments on the adding,<br />
removal <strong>and</strong> crossing out <strong>of</strong> tasks, <strong>and</strong> laughs while<br />
explaining that it isnt that funny when things get<br />
added.<br />
On the page <strong>of</strong> May 9. a list <strong>of</strong> team members<br />
names followed by their initials have been added at the<br />
top <strong>of</strong> the list. In subsequent page revisions, the tasks<br />
become appended with initials. TB explains that this is<br />
about distribution <strong>of</strong> responsibility.<br />
As the interviewer browses more page versions, it<br />
is apparent that there are many versions from May 9.<br />
TB says that maybe the team at the time had some<br />
mock-arrangement to h<strong>and</strong> in. He explains that their<br />
supervisor liked to have the team make deliveries <strong>of</strong><br />
preliminary versions to him before the big deadlines,<br />
<strong>and</strong> that the team appreciated this because there was<br />
an extreme time squeeze towards the end when we had<br />
to <strong>work</strong> on the project, the code, <strong>and</strong> we had as good<br />
as finished the report then. And that was very good.<br />
Or else we would not have been able to finish. <strong>The</strong><br />
interviewer asks if this means that their supervisor had<br />
stressed them a lot to get going with the report. TB<br />
confirms this <strong>and</strong> says that he would in fact<br />
recommend to supervisors to stress the h<strong>and</strong>ing in <strong>of</strong><br />
5<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
155
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
intermediary reports. Also, he would recommend that<br />
teams conduct mock presentations.<br />
This example shows how the wiki documents<br />
project management practice (e.g. the active use <strong>of</strong> todo-lists)<br />
<strong>and</strong> the status <strong>of</strong> task completion over time.<br />
<strong>The</strong> distribution <strong>of</strong> responsibility indicates how the<br />
management <strong>of</strong> remaining tasks became more<br />
meticulous towards project deadline. Some feelings<br />
associated with this process <strong>and</strong> phase are reflected<br />
upon by TB: the frustration <strong>of</strong> having to add new tasks<br />
at a point when one would wish to see the list shrink.<br />
<strong>The</strong> number <strong>of</strong> revisions <strong>of</strong> the page on a particular<br />
date, <strong>and</strong> the amount <strong>of</strong> changes done then, reflect to<br />
the team member that there was something particular<br />
going on. He reconstructs (rather than recalls) that<br />
most likely there was an intermediate delivery going<br />
on, explaining this as an established collaboration<br />
practice between the team <strong>and</strong> their supervisor. TB<br />
reflects on the adequacy <strong>of</strong> this practice, arguing why it<br />
was beneficial to the project. He also draws the<br />
perspective further, generalizing to supervised project<br />
<strong>work</strong> (e.g. in the course) when recommending<br />
supervisors to require intermediate, mock-up deliveries<br />
to help project teams complete important <strong>work</strong> on time.<br />
4.4. Example 4: Project <strong>work</strong> during final<br />
spurt<br />
TB <strong>and</strong> the interviewer are looking at the main<br />
page from May 23. This is the very last phase <strong>of</strong> the<br />
project, two days before delivery. <strong>The</strong>re is an item on<br />
top <strong>of</strong> the News bulletin list, with a link: Smarty stuff<br />
contains some Smarty functionality to be used in our<br />
templates. <strong>The</strong> interviewer asks what this is. TB starts<br />
explaining what he thinks it means, but then<br />
remembers <strong>and</strong> corrects himself: TB: It is in fact.. It<br />
was the last night, that one. When we sat fixing the<br />
code. And added functionality that we had not had the<br />
time for yet. So it is really just.. things that should have<br />
been there some months before. To put it that way. Not<br />
a day before. But what to do<br />
In the page version from 04:41 on the same night,<br />
the item Sudden realizations has been included in the<br />
News bulletin, with the following sub item:<br />
person_event.paid is currently not considered! How<br />
would this fit into our design at all?!. <strong>The</strong> interviewer<br />
asks a bout this, <strong>and</strong> TB explains that it refers to a very<br />
important thing that they had forgotten. But luckily<br />
they managed to fix it on that night. <strong>The</strong> interviewer<br />
asks if they were sitting together in the team at the<br />
time. TB says yes, but then corrects himself based on<br />
the time <strong>of</strong> the page revision: No, that sudden<br />
realizations, that, lets see, 04:41, it must have been<br />
me remembering No! Shit! before I was about to go<br />
to bed, <strong>and</strong> then I went into the wiki quickly, because I<br />
had to remember it for the meeting the next day. <br />
So I guess theres no more activity that night.. <strong>The</strong><br />
interviewer goes on to the next revision <strong>of</strong> the page, in<br />
which another bullet point has been added: TB laughs<br />
<strong>and</strong> says that that, too, was remembered suddenly <strong>and</strong><br />
later fixed.<br />
On the page version <strong>of</strong> May 23, 22:22, even more<br />
Sudden realizations have been added. TB comments<br />
to the page:Yes, now we sit, here we are<br />
..21:56.. B:Yes, right, there are the things you<br />
remembered, now you have had a meeting.<br />
TB:Yeseh.. looks like it.. [].<br />
In this example, events during the last night <strong>and</strong><br />
day <strong>of</strong> the project are reconstructed. TB reflects about<br />
<strong>work</strong> that should have been done earlier, on stress,<br />
effort <strong>and</strong> luck. He uses detailed information from the<br />
wiki page (exact time <strong>of</strong> the revision) to reconstruct<br />
how he used the wiki at the time: as a means for<br />
capturing his ideas in the early morning hours (in his<br />
bed, from his home), to remember for the next days<br />
team meeting. TB thus accounts for the <strong>work</strong> practices<br />
in the project under the extreme conditions <strong>of</strong> final<br />
deadline approaching. Trying to reconstruct this<br />
exceptional phase, TB actively uses the time<br />
information on the page revisions. Occasionally the<br />
information requires some thinking to make sense.<br />
(E.g.,May 23, 04:41 is the <strong>work</strong>day/night before May<br />
23, 22:22).<br />
4.5. What can be seen from the examples?<br />
<strong>The</strong> findings show that at project team members<br />
facilitated traversal <strong>of</strong> a project wiki leads to recall <strong>of</strong>,<br />
<strong>and</strong> <strong>reflection</strong> on the project process. <strong>The</strong> wiki does not<br />
contain a coherent account <strong>of</strong> the process, but the<br />
contents are rich enough for the project manager to<br />
reconstruct a project trajectory piece by piece.<br />
Information in the wiki triggering recall <strong>and</strong><br />
<strong>reflection</strong> include time data (date <strong>and</strong> time <strong>of</strong> day <strong>of</strong> a<br />
revision, sequence <strong>of</strong> revisions, frequency <strong>of</strong><br />
revisions), authorship, page content, <strong>and</strong> page<br />
appearance.<br />
Issues being recalled <strong>and</strong> reflected upon include<br />
particular project events, project phases, collaboration<br />
within the team <strong>and</strong> with other stakeholders (e.g. roles<br />
<strong>and</strong> responsibilities, <strong>and</strong> the use <strong>of</strong> collaboration<br />
technology).<br />
Some more specific observations are:<br />
− Sequences <strong>of</strong> events are reconstructed with the aid<br />
<strong>of</strong> information about the time <strong>of</strong> each revision.<br />
− Comments <strong>and</strong> <strong>reflection</strong> about the use <strong>of</strong> the wiki<br />
in the project <strong>of</strong>ten leads to <strong>reflection</strong> on the project<br />
more generally, e.g. events, phases, practices <strong>and</strong><br />
collaboration structures.<br />
6<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
156
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
− <strong>The</strong> development <strong>of</strong> to-do-lists over time is a good<br />
source <strong>of</strong> information on project management.<br />
− In periods in which the density <strong>of</strong> wiki revisions is<br />
high, the changes provide a picture <strong>of</strong> the there <strong>and</strong><br />
then experience, as in the final project spurt.<br />
− Facts as well as feelings are addressed in the review.<br />
Contents that appear to have been emotionally laden<br />
as they were added to the wiki (e.g. the red, boldface<br />
<strong>and</strong> growing urgent message), evokes<br />
emotional reaction (in our example: laughter) in the<br />
interview situation.<br />
− Reflection may extend to SE project <strong>work</strong> at large,<br />
e.g. lessons learned <strong>and</strong> suggested guidelines for<br />
projects or project courses.<br />
5. Benefits <strong>of</strong> reviewing project wiki<br />
history in postmortem reviews<br />
Having seen how a postmortem interview can be<br />
aided by the use <strong>of</strong> a project wiki, we now return to the<br />
selected challenges <strong>of</strong> postmortem <strong>reflection</strong> listed in<br />
Section 3.2. We will have a look at each challenge in<br />
turn to see how it can be addressed through suitable<br />
use <strong>of</strong> a project wiki. In doing so, we propose some<br />
wiki functionality that would be useful in this respect.<br />
5.1. Using project wikis to integrate with<br />
existing <strong>learning</strong> systems<br />
<strong>The</strong> project wiki itself as used in our case study can<br />
be considered a part <strong>of</strong> the organizations <strong>learning</strong><br />
system. <strong>The</strong> information about the project process<br />
found in the wiki includes elements <strong>of</strong> evaluation <strong>of</strong><br />
the process itself, formally or informally produced <strong>and</strong><br />
conveyed. Examples are project status reports <strong>and</strong><br />
emotional expressions <strong>of</strong> state-<strong>of</strong>-affairs. Later use <strong>of</strong><br />
this information in a postmortem review is a way <strong>of</strong><br />
turning to account <strong>learning</strong> that has already taken<br />
place, developing insights further as they are viewed in<br />
hindsight <strong>and</strong> by different team members.<br />
In addition to utilizing the history resulting from<br />
day-to-day use <strong>of</strong> the project wiki, it is possible to use<br />
the wiki to provide closer integration between the <strong>work</strong><br />
process <strong>and</strong> the identification <strong>of</strong> issues important to<br />
underst<strong>and</strong> <strong>and</strong> reflect upon the process. A quick <strong>and</strong><br />
easy way <strong>of</strong> tagging <strong>and</strong>/or annotating wiki contents<br />
relevant to a particular issue would be useful in this<br />
respect. To capture history, the tagging should apply to<br />
page revisions <strong>and</strong> possibly sections <strong>of</strong> page revisions.<br />
For a postmortem review, it should be possible to<br />
generate chronological views <strong>of</strong> the tagged contents to<br />
support the reconstruction <strong>of</strong> a trajectory.<br />
To support <strong>learning</strong> across projects in an<br />
organization, the result <strong>of</strong> a postmortem review could<br />
be stored in the project wiki itself, if the wiki is kept as<br />
a knowledge repository about the project. <strong>The</strong><br />
postmortem review report in the wiki should contain<br />
links to wiki elements (page/section revisions) deemed<br />
particularly interesting to the project trajectory.<br />
5.2. Using project wikis to avoid<br />
oversimplification <strong>of</strong> complex issues<br />
Whereas a wiki page revision represents a snapshot<br />
in time, showing only a particular view <strong>of</strong> some<br />
aspects <strong>of</strong> the project, it simultaneously represents a<br />
recognizable scene from a complex <strong>work</strong> setting. <strong>The</strong><br />
associations following the encounter with a familiar<br />
surrounding may help the team recall events <strong>and</strong><br />
emotions <strong>and</strong> the meaning they were given at the time,<br />
as shown in our case study. Also, links in the tool may<br />
be followed between different pages <strong>and</strong> back <strong>and</strong><br />
forth in time, thus allowing for the exploration <strong>of</strong> how<br />
events are connected, e.g. causalities. In this way, it<br />
may be possible to reconstruct a more complex picture<br />
than what is possible based on memory-based recall.<br />
<strong>The</strong> proposed tagging <strong>of</strong> contents by team<br />
participants <strong>and</strong> chronological traversal <strong>of</strong> the marked<br />
contents may be used to ensure that issues important to<br />
the team are actually addressed in the wiki-aided<br />
postmortem review. <strong>The</strong> opportunity to individually<br />
tag information during day-to-day <strong>work</strong> may help in<br />
ensuring that everyones voice is heard in the common<br />
postmortem review. All team members active<br />
participation in the tagging is necessary for the<br />
approach to succeed in capturing the full complexity <strong>of</strong><br />
the project.<br />
5.3. Using project wikis to help avoid<br />
information overload<br />
<strong>The</strong> organization <strong>of</strong> wiki contents across pages <strong>and</strong><br />
revisions can help team members access <strong>and</strong> sort out<br />
important information <strong>and</strong> structure their review <strong>of</strong> that<br />
information. <strong>The</strong> focus on the wiki main page in the<br />
postmortem interview <strong>of</strong> our case is an example,<br />
showing a simple way to gain some focus in the review<br />
process. Also, the example shows that allowing the<br />
team member to explore promising tracks as they turn<br />
up during the review, is a good way <strong>of</strong> spurring further<br />
investigation <strong>of</strong> the process.<br />
On the other h<strong>and</strong>, given the context <strong>of</strong> a<br />
postmortem review, some structure might be needed<br />
for purposes <strong>of</strong> efficiency in terms <strong>of</strong> covering key<br />
issues in a limited amount <strong>of</strong> time [23]. <strong>The</strong>re are many<br />
ways <strong>of</strong> approaching historical data on a project, <strong>and</strong> it<br />
may be desirable to focus on different aspects <strong>of</strong> the<br />
project in turn (e.g. individual team members<br />
7<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
157
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
perspectives, team-customer relationship, development<br />
<strong>of</strong> requirements.) Again, team participation in the<br />
tagging <strong>of</strong> contents, whether in the there-<strong>and</strong>-then<br />
<strong>work</strong> situation or in reflective hindsight, would help<br />
ensuring that information considered important by<br />
those involved is given priority.<br />
6. <strong>The</strong> wiki walkthrough tool: reviewing<br />
selected aspects <strong>of</strong> project history<br />
In Section 5 we have argued that user-defined<br />
tagging <strong>of</strong> wiki page revisions combined with the<br />
possibility to review the set <strong>of</strong> tagged revisions, could<br />
make a project wiki even more useful for postmortem<br />
reviews. In this section, we demonstrate how a wiki<br />
can be extended in this way. Our wiki walkthrough<br />
tool (WWT) has been implemented as a prototype<br />
extension <strong>of</strong> DokuWiki. An elaborate presentation <strong>of</strong><br />
the tool is beyond the scope <strong>of</strong> the paper, but we<br />
provide a brief use scenario <strong>and</strong> a description <strong>of</strong> the<br />
tool functionality.<br />
6.1. A scenario for the use <strong>of</strong> a wiki<br />
walkthrough tool<br />
<strong>The</strong> following scenario is intended to show in a<br />
compact way how a project wiki can be used for<br />
postmortem <strong>reflection</strong>, aided by some extended wiki<br />
functionality.<br />
<strong>The</strong> context for the scenario is as follows:<br />
<strong>The</strong>re is a SE student project team with five<br />
members doing development <strong>work</strong> for a customer. <strong>The</strong><br />
team has a project wiki in which they include or link in<br />
artifacts related to the project process <strong>and</strong> product.<br />
<strong>The</strong> wiki is actively used in project management:<br />
Various to-do-lists are updated by all team members,<br />
<strong>and</strong> there is a news bulletin mainly updated by the<br />
project manager. <strong>The</strong> chapters in the project report<br />
are included in the wiki <strong>and</strong> converted to a coherent,<br />
printable report at need. In addition to the wiki, the<br />
team makes use <strong>of</strong> development tools, including a<br />
versioning system, <strong>and</strong> collaboration technology such<br />
as a shared calendar <strong>and</strong> instant messaging.<br />
<strong>The</strong> postmortem review scenario:<br />
Throughout the project every team member tags elements<br />
in the project wiki reflecting what she considers to be<br />
important events. <strong>The</strong> team member decides if the tag is to be<br />
visible to other team members. <strong>The</strong> following postmortem<br />
review takes place twice: just before the delivery <strong>of</strong> a<br />
preliminary project report, <strong>and</strong> before delivery <strong>of</strong> the final<br />
report:<br />
Each team member spends an hour going through her<br />
tagged wiki contents, modifying the trajectory made up <strong>of</strong><br />
those elements by visiting wiki pages <strong>and</strong> adding or removing<br />
tags. In a common <strong>work</strong>shop supervised by a project-external<br />
facilitator, each team member gets 15 minutes to present her<br />
trajectory, using a projector in the meeting room. <strong>The</strong> team<br />
next discusses the project <strong>and</strong> how their respective<br />
underst<strong>and</strong>ings differ.<br />
After the mid-term postmortem review, results <strong>of</strong> the<br />
discussion (e.g. conflicting viewpoints, new insights) are<br />
documented <strong>and</strong> actively used in project planning (e.g.<br />
changed priorities, changed roles) After the final postmortem<br />
review, the team decides on some elements in the wiki that<br />
should be included in their common documentation <strong>of</strong> the<br />
project process.<br />
<strong>The</strong> postmortem review in the scenario is designed<br />
to reflect acknowledged practice [23], e.g. encouraging<br />
individual recall <strong>and</strong> <strong>reflection</strong> followed by plenary<br />
presentation <strong>and</strong> discussion, <strong>and</strong> the creation <strong>of</strong> a<br />
project management report which is later actively used<br />
to achieve improvement in the project.<br />
<strong>The</strong> scenario has been evaluated by two expert<br />
groups experienced in the supervisor <strong>and</strong> customer<br />
roles in the projects <strong>of</strong> the course described in Section<br />
3. Here, we only report the main conclusions from the<br />
evaluation: <strong>The</strong> expert groups were positive about the<br />
potential <strong>of</strong> wiki history to be utilized in postmortem<br />
<strong>reflection</strong> on the project process. <strong>The</strong>y were also<br />
positive about the potential <strong>of</strong> the WWT to help a team<br />
achieve a more focused review.<br />
6.2. <strong>The</strong> wiki walkthrough tool functionality<br />
A prototype <strong>of</strong> the WWT was developed during<br />
spring 2008. In providing a brief description <strong>of</strong> the tool<br />
functionality in what follows, we take two screenshots<br />
as our starting point.<br />
Figure 3. Tagging a content page with an existing<br />
tag ('todo_historikk')<br />
Figure 3 shows a page from a project wiki. This is a<br />
DokuWiki into which the WWT functionality has been<br />
integrated. <strong>The</strong> user is browsing the wiki <strong>and</strong> decides<br />
that she wants to tag the particular page revision. She<br />
has clicked the Add tags button at the top <strong>of</strong> the page.<br />
In the pop-up window appearing in the middle <strong>of</strong> the<br />
page, the user gets the option <strong>of</strong> using an existing tag<br />
or creating a new one. She picks an existing one,<br />
todo_historikk, from the list. When she clicks Save,<br />
8<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
158
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
the page revision will be tagged with todo_historikk.<br />
While in the example it is the current revision <strong>of</strong> the<br />
page being tagged, the procedure would have been the<br />
same if the user was viewing an older revision <strong>of</strong> the<br />
page <strong>and</strong> wanted to tag it.<br />
When the user chooses All tags or Tag search at<br />
the upper right <strong>of</strong> a contents page like the one in Figure<br />
3, she can proceed to choose a tag <strong>and</strong> get a report<br />
showing the page revisions tagged by it. In Figure 4,<br />
the report shows contents tagged with todo_historikk<br />
In the report view, for each page, tagged revisions<br />
are shown together chronologically. <strong>The</strong> user may<br />
view one or more <strong>of</strong> the revisions. In Figure 4, three<br />
wiki pages have been tagged with todo_historikk.<br />
Start has three revisions tagged, but the user has<br />
chosen to view only the last one. For the page<br />
TODO, the user has chosen to view two <strong>of</strong> the tagged<br />
revisions. In the small window on top <strong>of</strong> the upper<br />
TODO revision, visibility <strong>of</strong> the revisions can be<br />
changed. <strong>The</strong> third page displayed in this report is the<br />
presentasjon_final page revision tagged in Figure 3.<br />
Figure 4. Report view showing wiki contents tagged<br />
with 'todo_historikk'<br />
In the report view there is an option <strong>of</strong> exp<strong>and</strong>ed or<br />
compact views <strong>of</strong> the pages. <strong>The</strong> user may go directly<br />
from to a page revision to explore it further, e.g. follow<br />
one <strong>of</strong> its links. She may later return to the report view.<br />
Also, from the report view, the report may be printed.<br />
<strong>The</strong> tool has not yet been subject to evaluation<br />
through full-fledged use in a real project.<br />
6.3. Related <strong>work</strong> on wiki tagging<br />
Tagging wiki contents is not a new idea as such.<br />
Many existing wiki extensions allow users to impose<br />
extra structure on a wiki in order to make it useful in<br />
new ways.<br />
<strong>The</strong> use <strong>of</strong> semantic web enabled technologies to<br />
improve navigation <strong>and</strong> search in a wiki, so called<br />
social tagging, has been implemented in SweetWiki<br />
[24] <strong>and</strong> Semantic MediaWiki [25]. Gnowsis [26], a<br />
semantic desktop including a wiki, implements the idea<br />
<strong>of</strong> collecting relevant information to reflect a<br />
personalized view. <strong>The</strong> use <strong>of</strong> structured tagging,<br />
including free, personal tagging, to provide<br />
personalized access to wiki information for purposes <strong>of</strong><br />
document management, is discussed in [27].<br />
History or chronology is not in particular focus in<br />
the aforementioned systems. However, whereas the<br />
opportunities to tag earlier page revisions are only<br />
partially addressed in the associated literature, we<br />
assume that some existing tagging tools largely cover<br />
the functionality proposed for the WWT. Accordingly,<br />
from a technical point <strong>of</strong> view, some existing tool<br />
could be used for our purposes, maybe after some<br />
minor tailoring. From a usability perspective, we hold<br />
that alternative usage <strong>of</strong> a tool, perhaps involving only<br />
a subset <strong>of</strong> the available functionality <strong>and</strong> requiring<br />
substantial adjustments, might introduce too much<br />
noise to meet our high-level requirement for minimal<br />
intrusiveness into the <strong>work</strong> process in which tagging<br />
should be a background feature.<br />
A tool that deserves particular consideration for its<br />
related functionality is WikiTrails [28], designed to<br />
help build context <strong>and</strong> structure around existing wiki<br />
contents in an educational setting. <strong>The</strong> tool makes it<br />
possible for a wiki user to leave a trail while browsing<br />
the wiki, enabling other users to follow that trail later<br />
on. <strong>The</strong> aim is thus related to ours: building <strong>and</strong><br />
sharing a story through a sequence <strong>of</strong> wiki pages. A<br />
manual mode is provided, in which pages for a trail<br />
may be picked <strong>and</strong> added by the user, e.g. from a list.<br />
Two aspects <strong>of</strong> WikiTrails make it inadequate for<br />
the intended use <strong>of</strong> the WWT. First, WikiTrails is not<br />
designed to support the marking <strong>of</strong> potentially<br />
interesting information element during <strong>work</strong> activity<br />
for possible later incorporation in a trail. Second,<br />
WikiTrails provides no support for the generation <strong>of</strong> a<br />
chronological sequence; focus is on generating a trail<br />
that will amount to a pedagogically sound presentation<br />
<strong>of</strong> existing wiki contents. A third <strong>and</strong> minor objection<br />
to WikiTrails in our context is that is seems to lack<br />
functionality for the tagging <strong>of</strong> sections within a page,<br />
a drawback it shares with the WWT prototype.<br />
However, in WikiTrails the opportunity to annotate<br />
might compensate for this fairly coarse granularity.<br />
In sum, while acknowledging the potential <strong>of</strong><br />
existing tools to aid the structuring <strong>and</strong> traversal <strong>of</strong><br />
wiki contents, we advocate the use <strong>of</strong> a designated tool<br />
for the reconstruction <strong>of</strong> a project trajectory.<br />
9<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
159
Proceedings <strong>of</strong> the 42nd Hawaii International Conference on System Sciences - 2009<br />
7. Conclusion <strong>and</strong> further <strong>work</strong><br />
In this paper, we have drawn upon a detailed study<br />
<strong>of</strong> a postmortem interview from a capstone SE student<br />
project to demonstrate the potential <strong>of</strong> project wikis to<br />
be used as a means for the recall <strong>of</strong>, <strong>and</strong> <strong>reflection</strong> on,<br />
a project process, thus making the project wikis already<br />
in use even more useful to the project teams.<br />
Some limitations to our research provide directions<br />
for further research. <strong>The</strong> interview forming the core <strong>of</strong><br />
our empirical data was partially exploratory, <strong>and</strong><br />
adequate for pointing out the potential for project wikis<br />
to aid in postmortem <strong>reflection</strong>. Our findings indicate<br />
how various types <strong>of</strong> information may be used to recall<br />
<strong>and</strong> reflect on various types <strong>of</strong> issues. However, we do<br />
not have sufficient data to reasonably generalize about<br />
the connections between the two. We suggest that<br />
wiki-aided postmortem reviews be conducted in other<br />
projects <strong>and</strong> the interaction <strong>and</strong> tool use <strong>of</strong> the<br />
participants during the sessions be analyzed. <strong>The</strong><br />
interviews should follow a well founded design for a<br />
postmortem review <strong>and</strong> be video recorded. In this way<br />
it might be possible to identify what type <strong>of</strong><br />
information in a project wiki promotes <strong>reflection</strong> in<br />
certain ways or on certain topics. Also, collective, toolmediated<br />
postmortem <strong>reflection</strong> should be investigated.<br />
Our wiki walkthrough tool needs to be evaluated<br />
through use in real projects to see if it is deemed useful<br />
by the project teams.<br />
<strong>The</strong> idea <strong>of</strong> utilizing history stored in a<br />
collaboration tool in <strong>reflection</strong> on the process is<br />
relevant to other tools. E.g., source code management<br />
<strong>and</strong> bug tracking tools contain a history <strong>of</strong> commented<br />
project artifact revisions. <strong>The</strong> student teams at our<br />
university increasingly prefer this type <strong>of</strong> tool (e.g.<br />
Trac) to manage their SE projects. A next step in our<br />
research is to conduct <strong>and</strong> evaluate postmortem<br />
reviews mediated by these tools to see if they, too, can<br />
help students learn from their SE project experience.<br />
7. Acknowledgements<br />
Thanks to Monica Divitini, Arne Martin Aurlien,<br />
Håvard Håskjold <strong>and</strong> William Young. <strong>The</strong> research<br />
was funded by the MOTUS2 project at NTNU.<br />
8. References<br />
[1] P. H. Carstensen <strong>and</strong> K. Schmidt, "<strong>Computer</strong> Supported<br />
Cooperative Work: New Challenges to Systems Design," in<br />
H<strong>and</strong>book <strong>of</strong> Human Factors/Ergonomics, K. Itoh, Ed.<br />
Tokyo: Asakura Publishing, 2002 (1999).<br />
[2] G. Fitzpatrick, W. J. Tolone, <strong>and</strong> S. M. Kaplan, "Work,<br />
Locales <strong>and</strong> Distributed Social Worlds," presented at CSCW,<br />
Stockholm, Sweden, 1995.<br />
[3] M. Keil, J. Mann, <strong>and</strong> A. Rai, "Why S<strong>of</strong>tware Projects<br />
Escalate: An Empirical Analysis <strong>and</strong> Test <strong>of</strong> Four <strong>The</strong>oretical<br />
Models," MIS Quarterly, vol. 24, 2000.<br />
[4] J. Dewey, Democracy <strong>and</strong> education. New York: <strong>The</strong><br />
Free Press, 1997 (1916).<br />
[5] D. Schön, <strong>The</strong> Reflective Practitioner:Basic Books, 1983.<br />
[6] C. Argyris <strong>and</strong> D. Schön, Organizational Learning II.<br />
Reading, MA, USA: Addison Wesley, 1996.<br />
[7] P. Senge, <strong>The</strong> Fifth Discipline. London, United Kingdom:<br />
Doubleday, 1990.<br />
[8] E. Wenger, "Communities <strong>of</strong> practice <strong>and</strong> social <strong>learning</strong><br />
systems," Organization, vol. 7, pp. 225-246, 2000.<br />
[9] P. C. Blumenfeld, E. Soloway, R. W. Marx, J. S. Krajcik,<br />
M. Guzdial, <strong>and</strong> A. Palinscar, "Motivating Project-Based<br />
Learning: Sustaining the Doing, Supporting the Learning,"<br />
Educational Psychologist, vol. 26, pp. 369-398, 1991.<br />
[10] J. Dewey, How we think. A restatement <strong>of</strong> the relation<br />
<strong>of</strong> reflective thinking to the educative process (Revised edtn.)<br />
ed. Boston: D. C. Heath, 1933.<br />
[11] G. H. Mead, Mind, Self <strong>and</strong> Society. Chicago: <strong>The</strong><br />
University <strong>of</strong> Chicago Press, 1934.<br />
[12] A. Strauss, Continual permutations <strong>of</strong> action. New York:<br />
Aldine de Gruyter, 1993.<br />
[13] P. Berger <strong>and</strong> T. Luckmann, <strong>The</strong> Social Construction <strong>of</strong><br />
Reality. London: Penguin Books, 1966.<br />
[14] G. Fitzpatrick, <strong>The</strong> Locales Frame<strong>work</strong>: Underst<strong>and</strong>ing<br />
<strong>and</strong> Designing for Wicked Problems: Kluwer Academic<br />
Publishers, 2003.<br />
[15] J. Lave <strong>and</strong> E. Wenger, Situated Learning. Legitimate<br />
peripheral participation. Cambridge: University <strong>of</strong><br />
Cambridge Press., 1991.<br />
[16] I. Sommerville, S<strong>of</strong>tware Engineering, Sixth Edition:<br />
Addison-Wesley, 2001.<br />
[17] E. J. Conklin <strong>and</strong> K. B. Yakemovic, "A Process-<br />
Oriented Approach to Design Rationale," Human-<strong>Computer</strong><br />
Interaction, vol. 6, pp. 357-391, 1991.<br />
[18] A. H. Dutoit, R. McCall, I. Mistrik, <strong>and</strong> B. Paech,<br />
"Rationale Management in S<strong>of</strong>tware Engineering: Concepts<br />
<strong>and</strong> Techniques," in Rationale Management in S<strong>of</strong>tware<br />
Engineering, A. H. Dutoit, R. McCall, I. Mistrik, <strong>and</strong> B.<br />
Paech, Eds.: Springer, 2006.<br />
[19] T. Dingsøyr, "Postmortem reviews: purpose <strong>and</strong><br />
approaches in s<strong>of</strong>tware engineering," Information <strong>and</strong><br />
S<strong>of</strong>tware Technology, vol. 47, pp. 293-303, 2005.<br />
[20] B. Bygstad, B. Krogstie, <strong>and</strong> T.-M. Grønli, "Scaffolding<br />
Project Based Learning with the Rational Unified Process.<br />
Experience from 5 years <strong>of</strong> Student Projects in S<strong>of</strong>tware<br />
Engineering," presented at NOKOBIT, Norway, 2006.<br />
[21] V. Kasi, M. Keil, L. Mathiassen, <strong>and</strong> K. Pedersen, "<strong>The</strong><br />
post mortem paradox: a Delphi study <strong>of</strong> IT specialist<br />
perceptions," EJIS, vol. 17, pp. 62-78, 2008.<br />
[22] B. R. Krogstie, "<strong>The</strong> wiki as an integrative tool in<br />
project <strong>work</strong>," presented at COOP, Carry-le-Rouet,<br />
Provence, France, 2008.<br />
[23] S. M. Smith <strong>and</strong> E. Vela, "Environmental contextdependent<br />
memory: A review <strong>and</strong> meta-analysis,"<br />
Psychonomic Bulletin & Review, vol. 8, pp. 203-220, 2001.<br />
[24] "DokuWiki home page," vol. 2009.<br />
[25] E. Derby, D. Larsen, <strong>and</strong> K. Schwaber, Agile<br />
Retrospectives. Making Good Teams Great: Pragmatic<br />
Bookshelf, 2006.<br />
10<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:00 from IEEE Xplore. Restrictions apply.<br />
160
Research paper P6<br />
Title:<br />
Shared timeline <strong>and</strong> individual experience: Supporting retrospective <strong>reflection</strong> in<br />
student s<strong>of</strong>tware engineering teams<br />
Authors:<br />
Birgit Krogstie <strong>and</strong> Monica Divitini<br />
Published in:<br />
Proceedings <strong>of</strong> CSEE&T 2009<br />
Pages:<br />
8<br />
Complete reference:<br />
Krogstie, B. R. <strong>and</strong> M. Divitini (2009). Shared timeline <strong>and</strong> individual experience:<br />
Supporting retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering teams. Conference<br />
on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training (CSEE&T) 2009, Hyderabad, 17-19<br />
Feb. IEEE <strong>Computer</strong> Society.<br />
161
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
162
Shared timeline <strong>and</strong> individual experience: Supporting retrospective<br />
<strong>reflection</strong> in student s<strong>of</strong>tware engineering teams<br />
Birgit R. Krogstie <strong>and</strong> Monica Divitini<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology<br />
{birgitkr, divitini}@idi.ntnu.no<br />
Abstract<br />
To help SE student teams learn from their project process, we propose the use <strong>of</strong><br />
facilitated postmortem <strong>work</strong>shops in which each team reconstructs its project timeline.<br />
Individual team members’ experience <strong>of</strong> the project is included by team members drawing<br />
individual ‘experience curves’ along the timeline. <strong>The</strong> approach is based on current industry<br />
practice <strong>and</strong> adapted in accordance with theory on <strong>reflection</strong> <strong>and</strong> <strong>learning</strong>. We present the<br />
detailed design <strong>of</strong> the <strong>work</strong>shops as they were implemented in an undergraduate SE project<br />
course as well as some recommendations for future <strong>work</strong>shops. <strong>The</strong> <strong>work</strong>shops are effective<br />
for promoting active participation, constructing a shared representation <strong>of</strong> the project<br />
process, <strong>and</strong> revisiting project issues from multiple perspectives. Evaluation showed that<br />
students were satisfied with the <strong>work</strong>shop <strong>and</strong> motivated to use the approach in later projects.<br />
1. Introduction<br />
Learning from experience is essential in project <strong>work</strong>, whether in education or in industry.<br />
Such <strong>learning</strong> is about the day-to-day awareness <strong>of</strong> one’s own <strong>learning</strong> <strong>and</strong> about <strong>reflection</strong><br />
with some distance to events, i.e. retrospective <strong>reflection</strong>. In s<strong>of</strong>tware engineering (SE),<br />
retrospective <strong>reflection</strong> on the project process can be conducted in postmortems reviews, “a<br />
collective <strong>learning</strong> activity which can be organized for projects either when they end a phase<br />
or are terminated.” [1] p.295. <strong>The</strong> main purpose is to reflect on what happened in the project<br />
in order to improve future practice, for the individual participants <strong>and</strong> for the organization as<br />
a whole. A postmortem report is produced. <strong>The</strong> postmortem is typically facilitated.<br />
In industry SE projects, it is <strong>of</strong>ten a challenge to have teams prioritize postmortems [2],<br />
other tasks competing for resources. In student SE projects, reflecting on <strong>and</strong> writing about<br />
the process is not seen as ‘real’ project <strong>work</strong> (like coding <strong>and</strong> testing) <strong>and</strong> conflicts with other<br />
activities within or outside the project. Also, in a student team, it may be difficult to make<br />
everyone participate actively. <strong>The</strong> power hierarchy in the team (e.g. [2]) might impact on who<br />
is heard, who bothers to get involved, or who is allocated for the task. Thus, having a team do<br />
the process <strong>reflection</strong>, with full participation, is a challenge welcoming practical approaches.<br />
In student SE projects, successful postmortems should help students learn from the specific<br />
project <strong>and</strong> serve as a h<strong>and</strong>s-on introduction to st<strong>and</strong>ard SE practice. Further, from a course<br />
administration point <strong>of</strong> view, resource requirements (e.g. course staff time) must be feasible.<br />
To find a postmortem approach meeting these challenges, we look to theory on <strong>reflection</strong> <strong>and</strong><br />
<strong>learning</strong> <strong>and</strong> to current SE industry practice, as described in Section 2. Section 3 provides a<br />
step-by-step design <strong>of</strong> a postmortem <strong>work</strong>shop in an undergraduate SE course. Findings on<br />
students’ reflective process are presented in Section 4. In Section 5 we discuss the adequacy<br />
<strong>of</strong> our approach, making some recommendations. Section 7 concludes the paper.<br />
2. Background<br />
22nd Conference on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training<br />
Learning from project experience involves looking back at the process, identifying events,<br />
their sequence <strong>and</strong> their meaning. This may be considered as retrospectively reviewing the<br />
978-0-7695-3539-5/09 $25.00 © 2009 IEEE<br />
DOI 10.1109/CSEET.2009.20<br />
85<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
163
project trajectory <strong>and</strong> involves facts as well as feelings. Anselm Strauss [3] uses the term<br />
‘trajectory’ to describe: “(1) the course <strong>of</strong> any experienced phenomenon as it evolves over<br />
time” (e.g. an engineering project) <strong>and</strong> “(2) the actions <strong>and</strong> interactions contributing to its<br />
evolution” (pp.53-54). Events in a SE project can be understood in terms <strong>of</strong> the action <strong>and</strong><br />
interaction taking place within the team <strong>and</strong> between the team <strong>and</strong> other stakeholders. While<br />
Strauss uses the concept ‘arc <strong>of</strong> action’ to describe a trajectory as retrospectively understood,<br />
we apply the term ‘reconstructed trajectory’ to the retrospective underst<strong>and</strong>ing <strong>of</strong> the process.<br />
From a social constructivist perspective [4, 5], project reality is interpreted through the<br />
individual’s unique history <strong>of</strong> social interactions. Team members assign different meaning<br />
<strong>and</strong> significance to events based on their previous experience <strong>and</strong> can be expected to<br />
remember different facts as significant to the project. Making a shared, reconstructed<br />
trajectory can be seen creating common ground [6] for communicating about the project<br />
process. Possible disputes over facts may be settled by checking available evidence. In the<br />
case <strong>of</strong> the interpretation <strong>of</strong> events, completely shared underst<strong>and</strong>ing among team members is<br />
unattainable in principle, but participants may get a richer underst<strong>and</strong>ing <strong>of</strong> the process<br />
through the encounter with others’ perspectives. Contradictions or discrepancies between<br />
interpretations may serve as a source <strong>of</strong> reflective thinking [7]. To support reconstruction <strong>of</strong><br />
the shared trajectory <strong>and</strong> its use in <strong>reflection</strong>, it may be given a visual representation<br />
containing elements that are considered shared <strong>and</strong> agreed-upon as well as elements<br />
representing individual <strong>and</strong> possibly conflicting interpretations.<br />
Boud <strong>and</strong> colleagues present a model <strong>of</strong> <strong>reflection</strong> outlining three major components:<br />
experience(s), reflective process, <strong>and</strong> outcome [8]. Experience involves behavior, ideas <strong>and</strong><br />
feelings. In the reflective process addressing the experience, there are three steps: 1)<br />
Returning to experience, 2) attending to feelings, <strong>and</strong> 3) re-evaluating experience. Essentially,<br />
this means reconstructing the trajectory <strong>of</strong> the experience. Boud et al.’s model is adequate for<br />
guiding the reconstruction <strong>of</strong> a project trajectory in a postmortem review. In particular, the<br />
process <strong>of</strong> reconstruction can be approached by decomposing it into the steps <strong>of</strong> returning to<br />
experience, attending to feelings <strong>and</strong> re-evaluating experience. As Boud et al. stress, the steps<br />
should not be understood as a fixed template. Reflection is situated <strong>and</strong> creative, moving back<br />
<strong>and</strong> forth between steps <strong>and</strong> sometimes omitting them.<br />
Among the many SE postmortem techniques [1] there is an approach which reflects well<br />
the idea <strong>of</strong> reconstructing a project trajectory: the timeline exercise as outlined by Derby et al.<br />
[9]. An example <strong>of</strong> its use in industry can be found in [10]. <strong>The</strong> exercise is based on industry<br />
best practice for postmortems. While the context <strong>of</strong> [9] <strong>and</strong> [10] is agile development, the<br />
timeline technique is not bound to any particular s<strong>of</strong>tware life<strong>cycle</strong> model. Derby et al.<br />
outline a basic exercise with some possible variations. <strong>The</strong> purpose <strong>of</strong> the exercise is to<br />
stimulate memories, make a picture from many perspectives, examine assumptions about who<br />
did what when, <strong>and</strong> see patterns. <strong>The</strong> exercise may be used for ’just the facts’, or facts <strong>and</strong><br />
feelings.<br />
In the next section, we show how we have adapted <strong>and</strong> detailed the timeline exercise in<br />
line with the previously presented theory.<br />
3. Design <strong>of</strong> a postmortem <strong>work</strong>shop in a SE project course<br />
<strong>The</strong> case <strong>of</strong> our study is an undergraduate project course taking up 50% <strong>of</strong> students’<br />
<strong>work</strong>load in the last (6 th ) semester <strong>of</strong> the Bachelor program. <strong>The</strong> teams develop s<strong>of</strong>tware for<br />
genuine customers. Each team receives one grade <strong>and</strong> has a supervisor from course staff.<br />
Deliveries include a s<strong>of</strong>tware product, a project report in several versions <strong>and</strong> an oral<br />
86<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
164
presentation. In 2008, there were 11 teams <strong>of</strong> 3-5 students; 46 students altogether. <strong>The</strong> grade<br />
distribution in the course was 3 As, 4 Bs, 3 Cs <strong>and</strong> one D. B is a good grade that all but the<br />
most ambitious students would be satisfied with. <strong>The</strong> overarching <strong>learning</strong> goal <strong>of</strong> the course<br />
is to “get experience with SE project <strong>work</strong> in a team with a customer”, <strong>and</strong> postmortem<br />
<strong>reflection</strong> is seen as important to the <strong>learning</strong> outcome. In 2008 it was decided to provide<br />
support for this <strong>reflection</strong> by running a facilitated <strong>work</strong>shop with each team, to achieve active<br />
participation by all team members. <strong>The</strong> <strong>work</strong>shops were obligatory <strong>and</strong> scheduled between<br />
the final delivery <strong>and</strong> the oral project presentation.<br />
<strong>The</strong> author <strong>of</strong> this paper was the <strong>work</strong>shop facilitator. She started the semester as course<br />
coordinator, <strong>and</strong> after the course startup she continued only as researcher on the teams’ <strong>work</strong>.<br />
At the outset <strong>of</strong> the <strong>work</strong>shop, the students were informed that the facilitator was not involved<br />
in project evaluation, <strong>and</strong> that nothing said or written would be referred to outside the room,<br />
except made anonymous <strong>and</strong> for research purposes. All teams accepted audio recording <strong>and</strong><br />
having pictures taken <strong>of</strong> the resulting <strong>work</strong> on the whiteboard. <strong>The</strong> photos were sent to each<br />
team at the end <strong>of</strong> the <strong>work</strong>shop as a resource for their <strong>work</strong> on the <strong>reflection</strong> notes.<br />
<strong>The</strong> <strong>work</strong>shop setting was as follows: <strong>The</strong> participants were sitting by a table in a<br />
room with a large-size whiteboard. Each participant was provided with an A3 paper<br />
form containing a timeline marked with some major project events (e.g. main<br />
deliveries). On top <strong>of</strong> the sheet was a smiley face, <strong>and</strong> at the bottom a sad face (See<br />
Figure 1). On the table, there were pens <strong>and</strong> whiteboard markers in different colors.<br />
<strong>The</strong> <strong>work</strong>shop lasted 60 minutes <strong>and</strong> was divided into eight tasks (see Table 1). Task<br />
1 is the introduction, in which background information is provided. Tasks 2-3<br />
comprises individual <strong>and</strong> shared <strong>work</strong> to draw the project timeline. Individual timelines<br />
are made on the A3 sheets <strong>and</strong> the shared timeline is drawn by the facilitator on the<br />
whiteboard. Tasks 4-5 comprise individual drawing <strong>of</strong> experience curves on the A3<br />
sheets <strong>and</strong> the drawing <strong>and</strong> presentation <strong>of</strong> the curves along the timeline on the<br />
whiteboard (see Figure 1). With reference to Boud et al.’s <strong>reflection</strong> model, tasks 2-5<br />
can be seen as a ‘Returning to experience’ <strong>reflection</strong> step. Tasks 4-5 explicitly prescribe<br />
‘attending to feelings’. In tasks 6-7, the team is ‘re-evaluating experience’ by taking<br />
different perspectives: those <strong>of</strong> tasks, roles <strong>and</strong> lessons learned. In this way, issues that<br />
have emerged in tasks 2-5 are re-examined, <strong>and</strong> issues that have so far been missed may<br />
be brought up. <strong>The</strong> wrap-up in task 8 encourages an overall perspective on the process<br />
<strong>and</strong> can be used by the facilitator to have participants’ feedback on the <strong>work</strong>shop itself.<br />
Task name<br />
1 Intro<br />
(5 min)<br />
2 Individual<br />
timelines<br />
(5 min)<br />
3 Shared<br />
timeline<br />
(15 min)<br />
4 Individual<br />
experience<br />
curves<br />
(5 min)<br />
5 Present<br />
curves<br />
(10 min)<br />
Description<br />
Explain the purpose <strong>and</strong> agenda <strong>of</strong> the <strong>work</strong>shop. Clarify issues <strong>of</strong> confidentiality<br />
<strong>and</strong> research<br />
Each participant gets an A3 sheet <strong>of</strong> paper with a timeline reporting common<br />
events in the course (mainly the deliveries). <strong>The</strong> participants are asked to<br />
individually add events that they perceived had an impact on their project.<br />
Participants take turn in explaining the events they have listed <strong>The</strong> facilitator<br />
marks the events on the whiteboard on a timeline similar to the one on the<br />
individual sheets.<br />
<strong>The</strong> team members each draw their experience curve (or ‘satisfaction curve’) on<br />
the A3 sheet. <strong>The</strong> smiley face on top <strong>of</strong> the sheet indicates a level <strong>of</strong> great<br />
satisfaction. Down at the bottom is great dissatisfaction, <strong>and</strong> the timeline itself<br />
marks a neutral position in the middle.<br />
Each member in turn goes to the whiteboard, which holds the shared<br />
timeline. <strong>The</strong> team member first draws her curve with her whiteboard<br />
marker, next explains its shape. At the end <strong>of</strong> the session, all team<br />
members’ experience curves can be found on the whiteboard.<br />
87<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
165
6 Questions<br />
about roles<br />
& lessons<br />
learned<br />
(5 min)<br />
7 Present<br />
answers<br />
(10 min)<br />
8 Wrap-up<br />
(5 min)<br />
A sheet <strong>of</strong> incomplete statements addressing the project are uncovered,<br />
<strong>and</strong> the students are asked to turn their A3 sheet <strong>and</strong> write their answers on<br />
the blank page.<br />
1) “In the project, my role was…” 2) “Through the project, I got better at…”<br />
3) “In a similar project, I would like to become more skilled at…” 4) “<strong>The</strong><br />
most important thing I have learnt about s<strong>of</strong>tware engineering in this project<br />
is…”<br />
<strong>The</strong> answers to the questions are presented around the table<br />
<strong>The</strong> students may add further comments about their project on their sheet.<br />
If formal evaluation <strong>of</strong> the <strong>work</strong>shop is not done on another occasion, this is<br />
an opportunity to have feedback, oral or written. <strong>The</strong> A3 sheets are left for<br />
the facilitator/course staff.<br />
Table 1: Schedule <strong>of</strong> the postmortem <strong>work</strong>shop<br />
Our data collection is based on various observation data from the <strong>work</strong>shops, <strong>and</strong> on<br />
students’ course evaluation. <strong>The</strong> evaluation form, filled in by the students after the oral<br />
presentation, included two items on the <strong>work</strong>shop. Our observation data include the <strong>work</strong>shop<br />
audio recordings, some <strong>of</strong> which were summarized <strong>and</strong> partially transcribed. Excerpts<br />
were translated into English at need. <strong>The</strong> data also includes pictures <strong>of</strong> the <strong>work</strong>shop<br />
results on the whiteboard, <strong>and</strong> individual A3 sheets. <strong>The</strong> author’s knowledge <strong>of</strong> the course as<br />
researcher <strong>and</strong> staff is a backdrop for interpretation <strong>of</strong> the data. <strong>The</strong> whiteboard pictures were<br />
prepared for analysis <strong>and</strong> presentation using st<strong>and</strong>ard photo-editing s<strong>of</strong>tware, <strong>and</strong> were<br />
visually inspected <strong>and</strong> compared to look for observable patterns. Pictures were manipulated<br />
with felt pen on printouts to make the curves more visible.<br />
4. Findings<br />
In this section, we present our findings. We have selected two teams, coined P <strong>and</strong> Q, to<br />
provide some illustrative examples.<br />
Figure 1: Timeline <strong>and</strong> individual experience curves (originally in different colors) for team P.<br />
We have numbered the points along Max's curve where he explained the up- or downturn.<br />
4.1. Perspectives on the same process are different among team members<br />
Looking at the individual timelines within each team, we find that team members<br />
mention partially overlapping, but markedly different sets <strong>of</strong> events. This applies to all<br />
the teams in our study. Looking at the satisfaction curves within each team, we see that<br />
no teams have completely congruent curves. <strong>The</strong> curve sets range from ‘more or less<br />
88<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
166
congruent’ to sets that seem to originate in totally different projects. In 9 out <strong>of</strong> the 11<br />
projects, the curves appear to reflect very different experiences within the team. In<br />
Figure 1, for example, it can be seen that the student we have called Max has a curve<br />
with many turns, reflecting shifting spirits throughout the project. Ellis’ mood seems to<br />
have been fairly stable <strong>and</strong> positive. It can be seen from the curves that Ellis was the<br />
only team member who had a good time in the period just before the midterm report<br />
delivery (the timeline event by number 9). Sean <strong>and</strong> Max experienced a steep upturn in<br />
connection with the delivery. Finlay took longer to reach above-average spirits.<br />
We have numbered the points along Max’ explanation curve where he accounted for the<br />
up- or downturns. To illustrate the richness <strong>of</strong> an explanation <strong>of</strong> an individual curve, we<br />
provide Max’ explanation at the points 10, 11 <strong>and</strong> 12 (Figure 1):<br />
10) ”<strong>The</strong>n there was a lot <strong>of</strong> frustration because <strong>of</strong> the team member who disappeared, <strong>and</strong> I<br />
personally got completely p** <strong>of</strong>f [..] that we could not get in touch <strong>and</strong> he didn’t gave a damn.<br />
[…] We approached deadline, I felt that I was nagging like crazy <strong>and</strong> was rather stressed<br />
because I feared we would not reach our goal with all that needed to get finished ”. 11) ”<strong>The</strong>n<br />
we approached delivery <strong>and</strong> then I saw that this was going to go really well, <strong>and</strong> up went my<br />
spirits.”. 12) ”Spirit-wise I ended up approximately on Finlay’s level. We were rather happy<br />
when we had delivered.”<br />
Excerpt 1: A part <strong>of</strong> Max’ explanation <strong>of</strong> his curve. Team P.<br />
From this excerpt it can be noted that Max addresses his personal feelings about the<br />
challenges <strong>of</strong> organizing the project, in this case the issue <strong>of</strong> a member who quit the project<br />
without informing the team. Max also accounts for his own behavior at the time: he was<br />
project manager (not formally, but in practice) <strong>and</strong> pushed the other team members towards<br />
their deadline – which he perceived as ‘nagging’ <strong>and</strong> very stressful. He explains how the<br />
subsequent success <strong>of</strong> the project <strong>work</strong> affected him, comparing his final level <strong>of</strong> spirit with<br />
that <strong>of</strong> another team member. He also summarizes the team’s feeling at the end <strong>of</strong> the project.<br />
4.2. Making a shared timeline is more than a mere adding <strong>of</strong> individual timelines<br />
<strong>The</strong> construction <strong>of</strong> the shared timeline in task 3 <strong>of</strong>ten leads to discussion <strong>of</strong> details,<br />
significance, sequence <strong>and</strong> timing. When an event occurred in more than one timeline,<br />
discussion <strong>of</strong> the naming <strong>of</strong> the event <strong>of</strong>ten revealed discrepancies in the interpretation<br />
<strong>of</strong> elements <strong>of</strong> the project process. In many teams, events that had not emerged in task 2<br />
(drawing individual timelines) turned up in task 3. <strong>The</strong>se new events might be given as<br />
reasons for, or consequences <strong>of</strong>, other events.<br />
<strong>The</strong> most important finding from the shared timeline construction is that the<br />
facilitated ‘return to experience’ results in a shared representation different from the<br />
sum <strong>of</strong> individual timelines. Some events might be added or canceled, others renamed.<br />
In any case participants build common ground for <strong>reflection</strong> in the following tasks.<br />
4.3. <strong>The</strong> satisfaction curves give new insights about the project<br />
We have observed that the students glance at each other’s curves with interest in task<br />
4. Visible differences between curves appear to trigger puzzlement <strong>and</strong> amusement. In<br />
task 5, curve explanations in some cases trigger team members’ there-<strong>and</strong>-then response<br />
<strong>and</strong> <strong>reflection</strong>. In team Q, Magda explains a low part <strong>of</strong> her curve. She says she had<br />
been depressed about being assigned to a project using technology she already knew: “it<br />
was only project writing <strong>and</strong> project <strong>work</strong> that I learned anything about.” Peter supports<br />
<strong>and</strong> elaborates on Magda’s account, suggesting that it might be hard to be the only team<br />
89<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
167
member skilled in the technology. He thus acknowledges the view that improving<br />
technical skills is essential to a positive project experience.<br />
4.4 Through the <strong>work</strong>shop tasks, different perspectives are taken on important issues<br />
We see that the teams, by going through the 8 tasks, revisit issues <strong>of</strong> importance to<br />
their project. In task 7, answers frequently can be seen as ‘re-evaluating experience’,<br />
e.g. as major issues from the curve explanations are addressed in the lessons learned.<br />
<strong>The</strong> examination <strong>of</strong> issues from different perspectives sometimes reveals tensions<br />
<strong>and</strong> what appear as contradictions. <strong>The</strong>se add to the picture <strong>of</strong> the dynamics <strong>and</strong><br />
development <strong>of</strong> the project <strong>and</strong> fuel <strong>reflection</strong> on the process. An example from team P<br />
illustrates this. During task 5, Max explains that he was ‘nagging’ the team (Excerpt 1).<br />
Later, Sean explains a downturn in his curve during the same period: “Max was away<br />
for a week, <strong>and</strong> that was good for everyone. After that, it just went steadily upwards.“<br />
During task 7, explaining about roles in the project, Finlay says that he <strong>and</strong> Max, as an<br />
early bird <strong>and</strong> a B-person, respectively, had been quarreling a lot. Max says he had<br />
un<strong>of</strong>ficial roles like ‘A****le-Max’ <strong>and</strong> ‘Fusser-Max’. Answering what he had become<br />
better at, Max says “accepting that people are different”, e.g. in terms <strong>of</strong> preferred <strong>work</strong><br />
hours. About lessons learned, Sean says: “Don’t make conflicts that can be avoided”,<br />
explaining that sometimes, people had been “on the verge <strong>of</strong> psychotic outbursts”. His<br />
comment leads to discussion <strong>and</strong> agreement in the team that they had never let the sun<br />
go down on their quarrels, which were never personal. <strong>The</strong> facilitator refers to the<br />
whiteboard, asking if it reflects who had been quarreling the most, Sean looks at the<br />
curves (Figure 1), pauses, <strong>and</strong> says: “Yes, Finlay <strong>and</strong> Max” – which evokes laughter.<br />
During task 8, the team summarizes: Max: “ was awesome, really.”<br />
Finlay: “We reached our goal.” [..] Max:”It would have been worse if we had had a<br />
team member that we did not get along so well with” [..] ”We are largely similar, the<br />
whole bunch.” Ellis:”I personally could not have had it better with the team really.”<br />
4.5 Students perceive the <strong>work</strong>shop as useful<br />
Students’ perception <strong>of</strong> the insights gained through the <strong>work</strong>shop can be read from<br />
the course evaluation, in which agreement to the following two statements was rated: 1)<br />
“<strong>The</strong> postmortem <strong>work</strong>shop gave me insights about the project.” 2) “In another project<br />
in the future, I would suggest the use <strong>of</strong> timeline <strong>and</strong> experience curves to help the team<br />
reflect on the process”. 44 out <strong>of</strong> 46 students h<strong>and</strong>ed in the form, giving a 96% response<br />
rate. To statement 1, 2% <strong>of</strong> the students responded “strongly disagree” or “disagree”,<br />
25% responded “neither agree nor disagree”, <strong>and</strong> 73% responded “agree” or “strongly<br />
agree”. To statement 2, the respective percentages were 2%, 39% <strong>and</strong> 59%. From this,<br />
we may conclude that the students generally considered the postmortem approach<br />
useful to their project <strong>and</strong> worthwhile applying in another project.<br />
5. Discussion<br />
In this section, we discuss the relevance <strong>of</strong> our postmortem approach to SE education<br />
<strong>and</strong> the adequacy <strong>and</strong> possible improvements <strong>of</strong> the specific <strong>work</strong>shop design.<br />
5.1 Relevance to SE education<br />
Our findings suggest that the timeline <strong>and</strong> experience curve technique is adequate for<br />
postmortem <strong>reflection</strong> in SE student teams. Workshop observation indicated that the aspects<br />
<strong>of</strong> <strong>reflection</strong> outlined by Boud et al. are generally supported. <strong>The</strong> individual perspectives <strong>of</strong><br />
90<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
168
team members <strong>and</strong> the co-constructive effort <strong>of</strong> making a shared timeline results in a ‘return<br />
to experience’ richer than would have been possible for an individual, reflecting team<br />
member alone. Feelings are attended to in the drawing <strong>of</strong> satisfaction curves. Experience is<br />
re-evaluated through the revisiting <strong>of</strong> issues in the different <strong>work</strong>shop tasks, particularly the<br />
last ones. Further, the students deemed the <strong>work</strong>shop approach useful for later projects.<br />
Turning from student’s <strong>learning</strong> to organizational <strong>learning</strong>, three findings address the<br />
potential for course staff to use the <strong>work</strong>shops to learn about the course:<br />
First, students’ discussions provide insights about the effect <strong>of</strong> the course design. I.e.,<br />
the example from team Q in Section 4.3 points to different stakeholders having<br />
different project objectives [11]. Magda primarily wants to use the project to improve<br />
her technical skills, while the formal course objective is to have students get experience<br />
with project <strong>work</strong> for a customer, with a focus on the organization <strong>of</strong> team<strong>work</strong> <strong>and</strong><br />
management <strong>of</strong> stakeholder relationships. Such insights on students’ objectives <strong>and</strong><br />
motivation may be used to better align course objectives <strong>and</strong> actual <strong>learning</strong> outcome,<br />
e.g. by presenting the course differently or changing incentives or course requirements.<br />
Second, in our data, no patterns can be found relating curve sets to project grades.<br />
While data from a much larger number <strong>of</strong> teams might have revealed some significant<br />
correlation, a team’s curve set cannot be a diagnostic tool in project evaluation. <strong>The</strong><br />
curves reveal only part <strong>of</strong> a complex reality. Our curve sets may however be seen as<br />
disproving possible preconceptions about successful collaboration. For instance, a<br />
project receiving an A might have discrepant (Figure 1) or largely congruent curves<br />
(another team in our data). Rich information is needed to underst<strong>and</strong> the connection<br />
between experience curves <strong>and</strong> project result, e.g. why discrepancies might represent<br />
‘fruitful dynamics’ rather than subversive conflicts (cf. Section 4.4).<br />
Third, it is possible to find patterns relating satisfaction curves to elements <strong>of</strong> the course<br />
design. <strong>The</strong>se patterns can be interpreted in light <strong>of</strong> students’ explanations <strong>of</strong> their curves. We<br />
found two patterns in our data: a) Many students account for what can be denoted ‘report<br />
dips’: down-periods before report deliveries, associated with stress <strong>and</strong> comments like<br />
‘writing report is simply boring’. b) Feedback from the supervisor (e.g. report versions)<br />
has great impact on the team members’ experience. For instance, one <strong>of</strong> our<br />
supervisors, known to be ‘strict’, had been giving feedback making students in several<br />
teams very disappointed <strong>and</strong> even demotivated for some time. This did however not<br />
correlate with a negative grade. Patterns thus relating course events to individual<br />
experience – whether revealing surprising connections or providing empirical evidence <strong>of</strong><br />
prior assumptions - can be used as a resource for improvements to the course design.<br />
We see a potential for course staff to gain insights about the course from the <strong>work</strong>shops,<br />
but students’ <strong>reflection</strong> <strong>and</strong> <strong>learning</strong> might suffer if an additional agenda is introduced.<br />
Recommendation 1: If the <strong>work</strong>shop is used to learn about the course, agree with students that<br />
data will be anonymous <strong>and</strong> used for improvement <strong>of</strong> next year’s course. Recommendation 2:<br />
Make sure, <strong>and</strong> make students aware, that the <strong>work</strong>shop facilitator is not grading the project.<br />
5.1 Suitability <strong>of</strong> the proposed <strong>work</strong>shop structure<br />
Observations indicate that the <strong>work</strong>shop tasks were easily understood <strong>and</strong> willingly<br />
performed. <strong>The</strong>re were few questions, seemingly no misunderst<strong>and</strong>ings <strong>and</strong> no apparent<br />
reluctance to draw <strong>and</strong> explain curves on the whiteboard. All turns were explained <strong>and</strong><br />
linked to events, <strong>and</strong> the presenter was rarely interrupted. Explanations were generally<br />
given in a straightforward way, <strong>and</strong> the audience seemed sensitive to the information.<br />
91<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
169
Having the facilitator, <strong>and</strong> not the students, write down events on the timeline in task<br />
3 was time-efficient, but requires sensitiveness: the students should not be overrun. A<br />
facilitator in charge <strong>of</strong> the timeline gets an opportunity to support the reconstruction <strong>of</strong><br />
a shared timeline, e.g. when two students name the same event differently.<br />
Recommendation 3: use the proposed sequence <strong>of</strong> <strong>work</strong>shop steps. Shifting between<br />
individual <strong>and</strong> collective effort furthers active participation <strong>and</strong> multiple perspectives.<br />
We see that some teams might have benefited from a longer discussion. Need for<br />
more time typically increases with the size <strong>of</strong> the team. Recommendation 4: Use 90<br />
minute time slots, to have more flexibility to adjust to the specific team.<br />
In terms <strong>of</strong> feasibility for project courses <strong>of</strong> different sizes, the approach has a per-project<br />
cost <strong>of</strong> one staff hour plus preparation. Cost <strong>of</strong> preparation is likely to decrease with the<br />
number <strong>of</strong> teams facilitated by each staff member. While our <strong>work</strong>shop design is based on an<br />
agile development approach, it is adequate for other approaches. Recommendation 5: Use the<br />
<strong>reflection</strong> <strong>work</strong>shop as part <strong>of</strong> agile or more structured development approaches. Finally:<br />
Recommendation 6: If the <strong>work</strong>shop is to be conducted during the course <strong>of</strong> a project <strong>and</strong> not<br />
only at its end, the <strong>work</strong>shop should focus on measures for change, e.g. process improvement.<br />
6. Conclusion <strong>and</strong> further <strong>work</strong><br />
We have proposed a design for postmortem <strong>work</strong>shops in SE student teams, based on<br />
industry best practice <strong>and</strong> a theoretical frame<strong>work</strong> accounting for the <strong>reflection</strong> process.<br />
Based on its implementation in a SE course, we recommend our design, with additional<br />
recommendations made in Section 5. In our continued research <strong>and</strong> course development<br />
we will introduce a postmortem <strong>work</strong>shop in the middle <strong>of</strong> the projects, using a similar<br />
design with stronger focus on process improvement <strong>and</strong> specific change measures.<br />
7. Acknowledgements<br />
Thanks to the NTNU project MOTUS2, the SE students, Torgeir Dingsøyr,<strong>and</strong> John Krogstie.<br />
7. References<br />
[1] T. Dingsøyr, "Postmortem reviews: purpose <strong>and</strong> approaches in s<strong>of</strong>tware engineering," Information <strong>and</strong><br />
S<strong>of</strong>tware Technology, vol. 47, pp. 293-303, 2005.<br />
[2] V. Kasi, M. Keil, L. Mathiassen, <strong>and</strong> K. Pedersen, "<strong>The</strong> post mortem paradox: a Delphi study <strong>of</strong> IT specialist<br />
perceptions," European Journal <strong>of</strong> IS, vol. 17, pp. 62-78, 2008.<br />
[3] A. Strauss, Continual permutations <strong>of</strong> action. New York: Aldine de Gruyter, 1993.<br />
[4] P. Berger <strong>and</strong> T. Luckmann, <strong>The</strong> Social Construction <strong>of</strong> Reality. London: Penguin Books, 1966.<br />
[5] G. H. Mead, Mind, Self <strong>and</strong> Society. Chicago: <strong>The</strong> University <strong>of</strong> Chicago Press, 1934.<br />
[6] H. H. Clark <strong>and</strong> S. A. Brennan: Grounding in communication. In L.B. Resnick, J.M. Levine, & S.D. Teasley<br />
(Eds.). Perspectives on socially shared cognition . Washington: APA Books, 1991<br />
[7] J. Dewey, How we think. A restatement <strong>of</strong> the relation <strong>of</strong> reflective thinking to the educative process (Revised<br />
edtn.) ed. Boston: D. C. Heath, 1933.<br />
[8] D. Boud, R. Keogh, <strong>and</strong> D. Walker, Reflection: Turning Experience into Learning: RoutledgeFalmer, 1985.<br />
[9] E. Derby, D. Larsen, <strong>and</strong> K. Schwaber, Agile Retrospectives.: Pragmatic Bookshelf, 2006.<br />
[10] N. B. Moe <strong>and</strong> T. Dingsøyr, "Agile development <strong>and</strong> team<strong>work</strong>: A case study <strong>of</strong> a Scrum team <strong>and</strong> team<strong>work</strong><br />
organization," Work in progress, 2008.<br />
[11] B. Krogstie <strong>and</strong> B. Bygstad, "Cross-Community Collaboration <strong>and</strong> Learning in Customer-Driven S<strong>of</strong>tware<br />
Engineering Student Projects," in CSEE&T, Dublin, 2007, pp. 336-343.<br />
92<br />
Authorized licensed use limited to: Norges Teknisk-Naturvitenskapelige Universitet. Downloaded on February 5, 2010 at 09:01 from IEEE Xplore. Restrictions apply.<br />
170
Research paper P7<br />
Title:<br />
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning Utilizing Historical<br />
Data in Collaborative Tools<br />
Author:<br />
Birgit Krogstie<br />
Published in:<br />
Proceedings <strong>of</strong> EC-TEL 2009<br />
Pages:<br />
15<br />
Complete reference:<br />
Krogstie, B. R. (2009). A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong><br />
utilizing historical data in collaborative tools. European Conference on Technology<br />
Enhanced Learning (EC-TEL) 2009, Nice, France, 29 Sept – 2 Oct. Springer.<br />
171
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
172
A Model <strong>of</strong> Retrospective Reflection in Project Based<br />
Learning Utilizing Historical Data in Collaborative Tools<br />
Birgit R. Krogstie<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology (NTNU)<br />
birgitkr@gmail.com<br />
Abstract. In project based <strong>learning</strong>, <strong>learning</strong> from experience is vital <strong>and</strong> necessitates<br />
<strong>reflection</strong>. Retrospective <strong>reflection</strong> is as a conscious, collaborative effort<br />
to systematically re-examine a process in order to learn from it. In s<strong>of</strong>tware<br />
development student projects it has been empirically shown that project teams’<br />
retrospective <strong>reflection</strong> can help the teams collaboratively construct new<br />
knowledge about their process <strong>and</strong> that historical data in collaborative tools<br />
used in daily project <strong>work</strong> can aid the teams’ recall <strong>and</strong> <strong>reflection</strong> on the different<br />
aspects <strong>of</strong> project <strong>work</strong>. In this paper, we draw on these results as well as<br />
other findings on the use <strong>of</strong> collaborative tools in a similar setting. We use the<br />
frame<strong>work</strong> <strong>of</strong> distributed cognition to develop a model <strong>of</strong> retrospective <strong>reflection</strong><br />
in which collaborative tools used as cognitive tools for daily project <strong>work</strong><br />
are utilized as cognitive tools in retrospective <strong>reflection</strong>, aiding the creation <strong>of</strong><br />
individual <strong>and</strong> shared representations <strong>of</strong> the project process.<br />
Keywords: project based <strong>learning</strong>, <strong>reflection</strong>, distributed cognition, cognitive<br />
tools.<br />
1 Introduction<br />
In project based <strong>learning</strong> [1] <strong>learning</strong> from experience is essential [2-4]. To achieve<br />
this <strong>learning</strong>, <strong>reflection</strong> is necessary, both during day-to-day <strong>work</strong> <strong>and</strong> with some<br />
distance to the activity reflected upon [4]. In this paper, we will focus on what we will<br />
call retrospective <strong>reflection</strong>, a form <strong>of</strong> <strong>reflection</strong>-on-action in which participants in a<br />
collaborative process systematically re-examine their process.<br />
<strong>The</strong>re exist many techniques <strong>and</strong> tools to support <strong>reflection</strong> in project based <strong>learning</strong>.<br />
Examples include the writing <strong>of</strong> diaries <strong>and</strong> <strong>reflection</strong> notes, <strong>and</strong> user annotation<br />
<strong>of</strong> information in the collaborative tools used to support their <strong>work</strong>. <strong>Computer</strong>ized<br />
tools can be specifically introduced to support <strong>learning</strong>. Reflection supporting tools<br />
comprise tools showing users their <strong>learning</strong> process <strong>and</strong> its result, tools providing<br />
guidance for users’ monitoring <strong>of</strong> their <strong>learning</strong> process, <strong>and</strong> tools providing scaffolding<br />
for comparison with expert thinking [5]. In an organizational setting, advanced<br />
collaboration platforms may include knowledge management functionality [6], but<br />
such tools are untypical in formal education.<br />
<strong>The</strong> project based <strong>learning</strong> to be focused in this paper is that which is based on project<br />
<strong>work</strong> involving the development <strong>of</strong> artifacts <strong>and</strong> aided by lightweight collaborative<br />
U. Cress, V. Dimitrova, <strong>and</strong> M. Specht (Eds.): EC-TEL 2009, LNCS 5794, pp. 418–432, 2009.<br />
© Springer-Verlag Berlin Heidelberg 2009<br />
173
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 419<br />
tools. <strong>The</strong> latter includes tools frequently associated with Web 2.0, e.g. wikis, instant<br />
messaging tools, <strong>and</strong> discussion forums. Today’s students regularly use such tools to<br />
coordinate <strong>and</strong> perform their activities, whether imposed by school or not [7, 8]. Adding<br />
to the picture, students in higher education generally expect to have flexibility <strong>of</strong><br />
time <strong>and</strong> place for <strong>work</strong>. From the point <strong>of</strong> view <strong>of</strong> course organizers, lightweight tools<br />
are inexpensive options providing flexibility <strong>of</strong> use <strong>and</strong> functionality well known to<br />
students <strong>and</strong> course staff.<br />
A core argument in what follows is that by supporting various aspects <strong>of</strong> <strong>work</strong><br />
throughout a project, lightweight tools collect historical data with a potential value to<br />
help the students recall what happened in the project <strong>and</strong> reflect on it.<br />
<strong>The</strong> objective <strong>of</strong> the research in this paper is to provide a general account <strong>of</strong> retrospective<br />
<strong>reflection</strong> in the above described type <strong>of</strong> project based <strong>learning</strong> setting. We<br />
draw on empirical research on s<strong>of</strong>tware development (SD) student projects: mainly<br />
published <strong>work</strong> [9-11] <strong>and</strong> some additional findings. We provide a substantial new<br />
contribution by situating the results in a theoretical frame<strong>work</strong>, generalizing to project<br />
based <strong>learning</strong> independent <strong>of</strong> <strong>work</strong> domain <strong>and</strong> presenting a model which can be<br />
used to aid the organization <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong>.<br />
In Section 2, we outline the empirical research on SD student projects. We present<br />
our theoretical frame<strong>work</strong> in Section 3, <strong>and</strong> in Section 4 use it to analyze findings<br />
from the student projects. Section 5 outlines our conceptual model <strong>of</strong> retrospective<br />
<strong>reflection</strong>. Section 6 concludes the paper.<br />
2 Empirical Basis: Research on SD Student Projects<br />
In this paper we draw on findings from research on SD student project teams in the<br />
period 2006-2008 focusing on the teams’ use <strong>of</strong> collaboration technology. <strong>The</strong> project<br />
course is arranged in the last (6th) semester <strong>of</strong> an IT undergraduate program. Teams<br />
<strong>of</strong> 3-5 members develop s<strong>of</strong>tware for genuine customers. <strong>The</strong> main <strong>learning</strong> objective<br />
is to get experience with SD <strong>work</strong> in a team for a customer. Project deliveries include<br />
a s<strong>of</strong>tware product <strong>and</strong> a report. <strong>The</strong> teams choose which collaboration <strong>and</strong> development<br />
tools to use, depending on customer requirements, team members’ prior experience,<br />
team members’ wish to learn, <strong>and</strong> the availability <strong>of</strong> the technology.<br />
<strong>The</strong> data collection in the studies on these teams includes in-depth, longitudinal<br />
non-participant observation, semi-structured interviews across the cohorts, <strong>and</strong> examination<br />
<strong>of</strong> project documentation. A constructivist view <strong>of</strong> <strong>learning</strong> as well as<br />
guidelines for interpretive field research [12] guided the data collection <strong>and</strong> analysis.<br />
2.1 Retrospective Reflection Based on Participants’ Memory<br />
<strong>The</strong> adoption <strong>of</strong> a retrospective <strong>reflection</strong> technique from SD industry [13] in the SD<br />
project course was subject to a qualitative study [10]. Facilitated <strong>reflection</strong> <strong>work</strong>shops<br />
were conducted with 10 student teams at the end <strong>of</strong> their project. Each <strong>work</strong>shop<br />
lasted about an hour <strong>and</strong> started with participants’ individually drawing project timelines<br />
incorporating events they perceived as important to the project, <strong>and</strong> along the<br />
timelines, ‘satisfaction curves’ indicating their feelings about the project over<br />
time. From the individual timelines, the teams constructed shared timelines on a<br />
174
420 B.R. Krogstie<br />
Fig. 1. Shared timeline from a <strong>reflection</strong> <strong>work</strong>shop (One event emphasized by the authors)<br />
whiteboard. Individual satisfaction curves were drawn along it (see Fig. 1) <strong>and</strong> explained<br />
by the participant. Next, the participants individually answered questions<br />
about tasks, roles <strong>and</strong> lessons learned <strong>and</strong> presented their answers. After the <strong>work</strong>shop,<br />
the teams made <strong>reflection</strong> notes.<br />
<strong>The</strong> study showed that individual timelines <strong>and</strong> satisfaction curves reflected different<br />
perspectives on a project process. Fig. 1 illustrates how experience curves may<br />
differ within a team. <strong>The</strong> study also showed that shared timelines <strong>of</strong>ten reflected<br />
views <strong>of</strong> the project not found in any individual timeline. Closer examination <strong>of</strong> the<br />
individual timelines used in the creation <strong>of</strong> the shared timeline in Fig. 1 revealed that<br />
most <strong>of</strong> the events from the individual timelines had been included in the shared one,<br />
<strong>and</strong> some had been transformed through the co-constructive effort. For instance, the<br />
event ‘a bit ineffective <strong>work</strong>’ in an individual timeline was modified into an event<br />
marking a point in the project process when the team realized that they had to <strong>work</strong><br />
more efficiently (‘insight: need to <strong>work</strong> more efficiently’). <strong>The</strong> study [10] concludes<br />
that the satisfaction curves gave the students new insights, that the <strong>work</strong>shop helped<br />
them take new perspectives on important issues, <strong>and</strong> that they considered it useful.<br />
2.2 Retrospective Reflection Aided by Historical Data in Collaborative Tools<br />
In SD industry it is acknowledged [14] that project retrospectives would benefit from<br />
better data to help participants create a shared underst<strong>and</strong>ing while avoiding oversimplification<br />
<strong>and</strong> time-consuming examination <strong>of</strong> unimportant information. This was<br />
addressed in a study <strong>of</strong> SD student projects in which historical data in project wikis,<br />
used as lightweight project management tools by several teams [15], were used to aid<br />
project participants’ <strong>reflection</strong> on their project process [9]. In retrospective interviews<br />
with project members, wiki contents were chronologically examined, <strong>and</strong> particular<br />
types <strong>of</strong> information was seen to trigger recall <strong>of</strong> <strong>and</strong> <strong>reflection</strong> on project events,<br />
project phases, <strong>and</strong> collaboration within the team <strong>and</strong> with other stakeholders.<br />
In [11] it was shown that historical data in an issue tracking tool, a lightweight tool<br />
for SD project management, was useful to aid retrospective <strong>reflection</strong>. Historical<br />
data was accessed by traversal <strong>of</strong> a timeline showing team members’ updates to development<br />
artifacts. <strong>The</strong> <strong>reflection</strong> effort was organized in line with the approach <strong>of</strong><br />
175
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 421<br />
creating timelines <strong>and</strong> satisfaction curves [10], but conducted as a two day, video<br />
recorded <strong>work</strong>shop with one team to investigate in depth the use <strong>and</strong> role <strong>of</strong> historical<br />
data in the <strong>reflection</strong> process. Examination <strong>of</strong> the historical data helped team members<br />
individually identify project events that had not been included in the timelines made<br />
from (individual) memory alone <strong>and</strong> that were later included in the collaboratively<br />
constructed timeline <strong>and</strong> accounts <strong>of</strong> lessons learned. Historical data in the issue<br />
tracker were also used by the team to adjust details in the shared timeline.<br />
Fig. 2. Historical data in a collaborative tool aiding recall <strong>of</strong> a project event [11]<br />
Some features for navigation <strong>and</strong> retrieval <strong>of</strong> historical data were found to be important<br />
in retrospective <strong>reflection</strong> [9, 11]. <strong>The</strong>se features included chronological<br />
overview <strong>and</strong> traversal, <strong>and</strong> the possibility to switch between overview <strong>and</strong> details.<br />
Interviews with the 2007-2008 teams <strong>and</strong> examination <strong>of</strong> their <strong>reflection</strong> notes<br />
showed that the collaborative tools used were generally lightweight. <strong>The</strong> teams were<br />
clear about what tools were used for what purpose. While the specific tool selection<br />
<strong>and</strong> usage differ among teams, we see a general pattern:<br />
Lightweight project management tools (typically project wikis or issue tracking<br />
tools) are used for managing team-internal coordination <strong>of</strong> tasks, e.g. create <strong>and</strong> follow<br />
up on a project plan, <strong>and</strong> define, assign <strong>and</strong> track the status <strong>of</strong> tasks. <strong>The</strong> tool<br />
provides links to project documentation. Development tools are used to write, test <strong>and</strong><br />
integrate source code. <strong>The</strong> storage <strong>and</strong> versioning <strong>of</strong> project artifacts are managed in a<br />
file versioning system that may or may not be integrated with the project management<br />
tool. Email is used for formal <strong>and</strong> documented communication internally <strong>and</strong> in<br />
communication with other stakeholders. Typically, the project team has its own mailing<br />
list. Instant messaging chat is used for informal, team-internal messages <strong>and</strong> substitute<br />
face-to-face conversation over synchronous, distributed <strong>work</strong>. Less <strong>of</strong>ten, it is<br />
used for communication with other stakeholders. Internet sites are used to get<br />
176
422 B.R. Krogstie<br />
information about technology; in most cases, simple web search or FAQ lists provide<br />
answers, but occasionally project members participate in discussion forums [16].<br />
<strong>The</strong>se patterns <strong>of</strong> use, combined with the functionality <strong>of</strong> tools, imply that data resulting<br />
from various types <strong>of</strong> project <strong>work</strong> is generally logged. Wiki revisions are<br />
automatically stored <strong>and</strong> the email clients store all mail unless otherwise specified by<br />
the user. It is tacitly expected that an email user keeps an archive <strong>of</strong> <strong>work</strong>-related<br />
email. With instant messaging, team members frequently choose to enable logging.<br />
On an internet forum, postings remain as long as the community hosting the forum<br />
wants to keep them. Looking into the historical data stored in these tools, they can be<br />
seen as a trace <strong>of</strong> the <strong>work</strong> undertaken with the aid <strong>of</strong> the tools. <strong>The</strong> studies investigating<br />
<strong>reflection</strong> aided by historical data [9, 11] showed that data in tools used for project<br />
management <strong>and</strong> coordination reminded participants <strong>of</strong> events related to those aspects<br />
<strong>of</strong> project <strong>work</strong> <strong>and</strong> thus helped them reflect on that type <strong>of</strong> issues.<br />
To examine the potential for historical data to shed light on other types <strong>of</strong> issues,<br />
we started out by issues that, according to the teams’ own <strong>reflection</strong>. had been <strong>of</strong> great<br />
importance in their projects (e.g. misunderst<strong>and</strong>ings in team-customer collaboration,<br />
problems <strong>of</strong> getting timely information from a service provider). Examining historical<br />
data in collaborative tools used by those teams, including email archives, instant messaging<br />
logs, <strong>and</strong> discussion forums, we looked for historical data that could shed light<br />
on those issues <strong>The</strong> data found were, as seen by the researcher, rich enough to have<br />
such a potential, partially because the data shed light on the use <strong>of</strong> the collaborative<br />
tool itself, which was <strong>of</strong>ten at the core <strong>of</strong> the project challenge.<br />
We end this section by outlining some more characteristics <strong>of</strong> the SD projects relevant<br />
to our agenda <strong>of</strong> underst<strong>and</strong>ing <strong>and</strong> supporting retrospective <strong>reflection</strong> in the<br />
teams. Work in the projects is typically diversified, project participants having different<br />
roles, dividing <strong>work</strong> <strong>and</strong> using different tools to address different tasks. Team<br />
members’ roles affect their day-to-day use <strong>of</strong> collaborative tools. Consequently, historical<br />
data in a tool typically reflects <strong>work</strong> in which some team members have been<br />
more involved than others. Project artifacts (e.g. requirements specifications <strong>and</strong> project<br />
plans) frequently play a role in collaboration with project stakeholders (e.g.<br />
customer [17] <strong>and</strong> course staff) having different goals for their project involvement.<br />
Project artifacts in various states can be seen as more or less well defined versions.<br />
<strong>The</strong>se may be deliberately saved by the user (as when a text document is renamed <strong>and</strong><br />
saved) or automatically stored in a tool. A file versioning system stores the contents<br />
<strong>of</strong> <strong>and</strong> differences between every file version ‘checked in’ by the users. In our studies<br />
<strong>of</strong> retrospective <strong>reflection</strong> aided by historical data in collaborative tools, going into<br />
detail <strong>of</strong>ten meant exploring specific artifact versions.<br />
3 <strong>The</strong>oretical Background<br />
Taking the view <strong>of</strong> constructivism, seeing <strong>learning</strong> as integral to activity that is basically<br />
social <strong>and</strong> situated, <strong>and</strong> focusing on the role <strong>of</strong> tools, several theories [18, 19]<br />
may shed light on project based <strong>learning</strong>. <strong>The</strong>y include activity theory (AT), actor<br />
net<strong>work</strong> theory (ANT), symbolic interactionism, situated <strong>learning</strong>, <strong>and</strong> distributed<br />
cognition [20, 21]. In [11] it was shown that distributed cognition is an adequate<br />
frame<strong>work</strong> for underst<strong>and</strong>ing retrospective <strong>reflection</strong> in SD student projects. It has<br />
177
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 423<br />
been used to analyse <strong>learning</strong> in educational [22-24] as well as <strong>work</strong> [25, 26] settings.<br />
In the present study, the main reason for choosing distributed cognition among the<br />
c<strong>and</strong>idate frame<strong>work</strong>s is its focus on transformation between representations. Such<br />
transformations can be seen as a core element in a process <strong>of</strong> retrospective <strong>reflection</strong><br />
incorporating construction <strong>of</strong> timelines <strong>and</strong> examination <strong>of</strong> historical data. <strong>The</strong> concept<br />
<strong>of</strong> cognitive tools [23, 27] can shed light on how such representations aid <strong>work</strong><br />
<strong>and</strong> <strong>learning</strong>. Selecting distributed cognition as a frame<strong>work</strong> we focus on its descriptive<br />
power <strong>and</strong> what we want to achieve by applying it [18].<br />
We want to develop a model which not only descriptively accounts for the elements<br />
<strong>and</strong> dynamics <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> but which<br />
also informs its organization. To this end we augment our theoretical frame<strong>work</strong> with<br />
theory addressing how the process <strong>of</strong> <strong>learning</strong> <strong>and</strong> <strong>reflection</strong> may be facilitated.<br />
Reflection can be seen as active <strong>and</strong> careful consideration <strong>of</strong> any belief or supposed<br />
form <strong>of</strong> knowledge [28], implying the reviewing <strong>and</strong> judging <strong>of</strong> present knowledge or<br />
beliefs [29]. A model <strong>of</strong> the <strong>learning</strong> process linking <strong>reflection</strong> to the experience reflected<br />
on in a reciprocal relationship is provided by Kolb <strong>and</strong> Fry [30]. Boud et al.<br />
[31] present a model <strong>of</strong> the reflective process in which the returning to experience is<br />
essential. <strong>The</strong> experience comprises the behaviour, ideas <strong>and</strong> feelings involved in the<br />
situation reflected upon. In the reflective process, feelings about the experience<br />
should be attended to <strong>and</strong> the experience re-evaluated. <strong>The</strong> desired outcomes <strong>of</strong> <strong>reflection</strong><br />
include new perspectives on the experiences, change in behaviour, readiness<br />
for application <strong>and</strong> commitment to action. <strong>The</strong> model is descriptive with respect to<br />
steps typically occurring in <strong>reflection</strong> but is also meant to provide guidance about how<br />
to achieve successful <strong>reflection</strong>.<br />
Project <strong>work</strong> experience can be understood in terms <strong>of</strong> the project trajectory [32], a<br />
concept in line with the idea <strong>of</strong> temporarily distributed cognition. Strauss refers to<br />
Mead on the issue <strong>of</strong> how a social world reflects on its history: “<strong>The</strong> temporal spans<br />
<strong>of</strong> group life mean that the aims <strong>and</strong> aspirations <strong>of</strong> group endeavor are subject to reviewal<br />
<strong>and</strong> recasting. Likewise past activities come to be viewed in new lights,<br />
through reappraisal <strong>and</strong> selective recollection. ..History, whether that <strong>of</strong> a single person<br />
or <strong>of</strong> a group, signifies a ‘coming back at self’ (Mead 1936)”. By ‘trajectory’<br />
Strauss means “the course <strong>of</strong> any experienced phenomenon as it evolves over time”.<br />
Distributed cognition “implies that the learners are enabled to think deeply <strong>and</strong><br />
create certain types <strong>of</strong> artifacts that represent their thinking by <strong>work</strong>ing with cognitive<br />
tools” [23] p.209. Cognitive tools are tools that “help users represent what they know<br />
as they transform information into knowledge <strong>and</strong> are used to engage in, <strong>and</strong> facilitate,<br />
critical thinking <strong>and</strong> higher order <strong>learning</strong>.” [27]<br />
Stahl [33] explicitly addresses the collaborative aspect <strong>of</strong> <strong>learning</strong> in his model <strong>of</strong><br />
collaborative knowledge building, “a cyclical process with no beginning or end”<br />
(p.75). At the heart <strong>of</strong> the model are processes <strong>of</strong> building personal <strong>and</strong> collaborative<br />
knowing. <strong>The</strong> model comprises the transformation <strong>of</strong> cultural <strong>and</strong> cognitive artifacts<br />
<strong>and</strong> sheds light on the interplay <strong>of</strong> individual <strong>and</strong> collective <strong>reflection</strong> <strong>and</strong> <strong>learning</strong>.<br />
Our aim is to develop model that sheds light on the elements <strong>and</strong> dynamics <strong>of</strong> retrospective<br />
<strong>reflection</strong> in the particular setting <strong>of</strong> project based <strong>learning</strong> in which lightweight<br />
collaborative tools are used to support <strong>work</strong> <strong>and</strong> retrospective <strong>reflection</strong>. We<br />
want the model to account for the return to experience in the light <strong>of</strong> project trajectory<br />
178
424 B.R. Krogstie<br />
<strong>and</strong> to express how <strong>work</strong> <strong>and</strong> <strong>learning</strong> involves individual as well as collective <strong>reflection</strong><br />
<strong>and</strong> the use <strong>of</strong> cognitive tools.<br />
We clarify our use <strong>of</strong> some concepts: Activity is used as a generic, commonsense<br />
term <strong>and</strong> not as a reference to activity theory. By project artifact we mean something<br />
used <strong>and</strong> produced in project <strong>work</strong>, e.g. a diagram or a report. In distinguishing between<br />
<strong>work</strong> <strong>and</strong> the retrospective <strong>reflection</strong> on that <strong>work</strong>, we deliberately ‘hide’ the<br />
<strong>reflection</strong> <strong>and</strong> <strong>learning</strong> in day-to-day project <strong>work</strong> inside the concept <strong>of</strong> project <strong>work</strong>.<br />
This is not to pretend that project based <strong>learning</strong> only happens in ‘chunks’ <strong>of</strong> retrospective<br />
<strong>reflection</strong>, but it is our agenda to shed light on the latter.<br />
We proceed with an analysis <strong>of</strong> retrospective <strong>reflection</strong> on the above theoretical<br />
basis, from the empirical grounding described in Section 2 <strong>and</strong> with the objective <strong>of</strong><br />
developing a model. With a sidelong glance at the <strong>work</strong> on organizational memory in<br />
[25], we structure our analysis in terms <strong>of</strong> retrospective <strong>reflection</strong> being socially distributed,<br />
temporally distributed, <strong>and</strong> involving transformation <strong>of</strong> representations.<br />
4 Analysis<br />
We use findings from the research on SD teams outlined in Section 2 to shed light on<br />
how retrospective <strong>reflection</strong> may be supported in similar settings <strong>of</strong> project based<br />
<strong>learning</strong> from the perspective <strong>of</strong> distributed cognition.<br />
4.1 <strong>The</strong> Social Distribution <strong>of</strong> Cognition in Retrospective Reflection<br />
<strong>The</strong> social distribution <strong>of</strong> the cognition involved in retrospective <strong>reflection</strong> can be<br />
seen as having two main components: the social distribution <strong>of</strong> the process reflected<br />
upon, <strong>and</strong> that <strong>of</strong> the retrospective <strong>reflection</strong> activity itself. <strong>The</strong> social distribution <strong>of</strong><br />
the process reflected upon in the case <strong>of</strong> SD student projects was largely described in<br />
Section 2 <strong>and</strong> illustrates the complexity <strong>of</strong> the experience to be returned to in retrospective<br />
<strong>reflection</strong> We noted that tasks relating to different aspects <strong>of</strong> <strong>work</strong> were<br />
distributed in the teams, resulting in a distribution <strong>of</strong> tool use. Historical data were<br />
generally being stored in the tools as a result <strong>of</strong> the <strong>work</strong>. <strong>The</strong> data reflected aspects<br />
<strong>of</strong> project <strong>work</strong> in which the tool has played a role, including the tool use itself. E.g.,<br />
in Fig. 2, the historical data reflects s<strong>of</strong>tware development <strong>work</strong>, more specifically<br />
coding. Crucially, in our context, a new version <strong>of</strong> a project artifact stored in the tool<br />
represents different data than the previous version, this distinction serving to capture<br />
the temporal <strong>and</strong> partially also the social distribution <strong>of</strong> the project <strong>work</strong>.<br />
We now take a closer look at the social distribution <strong>of</strong> the retrospective <strong>reflection</strong><br />
with reference to the SD students teams. <strong>The</strong> timelines in [10] (see Fig.1) were drawn<br />
with the aid <strong>of</strong> simple physical tools; paper <strong>and</strong> pencil, whiteboard <strong>and</strong> pens. <strong>The</strong>se<br />
tools had a flexibility in the situation <strong>of</strong> knowledge sharing appearing to make them<br />
adequate for externalizing <strong>and</strong> transforming participants’ representations.<br />
When historical data in collaborative tools were introduced in similarly organized<br />
retrospective <strong>reflection</strong> [11], particular tool features for retrieving <strong>and</strong> navigating data<br />
were found to be important to the utility <strong>of</strong> the tool. For instance, the view in Fig.2<br />
allows easy chronological traversal with direct access to project artifacts in there-<strong>and</strong>then<br />
versions. <strong>The</strong> likelihood <strong>of</strong> being able to retrieve interesting data from a specific<br />
tool also depends on the actual tool usage in the team’s <strong>work</strong>. Further, what is<br />
179
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 425<br />
worthwhile examining retrospectively depends on what the team considers important<br />
issues. <strong>The</strong>se might have been identified prior to retrospective <strong>reflection</strong>, but may also<br />
emerge from the examination <strong>of</strong> the historical data as seen in [11] <strong>and</strong> Fig.2.<br />
Synchronous, distributed <strong>work</strong> among pairs <strong>of</strong> SD team members is frequently<br />
supported by instant messaging chat. Such conversation has an oral flavor, <strong>of</strong>ten<br />
combining there-<strong>and</strong>-then problem resolution with social chit-chat. Instant messaging<br />
logs exemplify historical data that for privacy reasons may be inadequate for retrospective<br />
<strong>reflection</strong> even if the contents are <strong>of</strong> potential interest to the team.<br />
<strong>The</strong> social distribution <strong>of</strong> retrospective <strong>reflection</strong> is also about which participants<br />
are involved, e.g. whether <strong>reflection</strong> is done individually or by the whole team<br />
together. In the SD teams, recall <strong>of</strong> events was essential in the construction <strong>of</strong> individual<br />
<strong>and</strong> shared timelines [10, 11]. Research on transactive memory, “a set <strong>of</strong> individual<br />
memory systems in combination with the communication that takes place between<br />
individuals” [34], shows that in collaborative remembering, comparing groups<br />
<strong>of</strong> people who have a history <strong>of</strong> remembering together (e.g. as colleagues or as family<br />
members) with groups people who have not, those who are used to repeatedly remembering<br />
together have the most efficient systems for doing so [35, 36]. In close<br />
relationships, responsibility for knowing is distributed, with the effect that some information<br />
is differentiated <strong>and</strong> non-redundant whereas other is shared. However,<br />
research indicates that the collective remembering skills that a group <strong>of</strong> people has<br />
developed through experience are only an advantage in a situation <strong>of</strong> remembering if<br />
the group is allowed to use these skills [37]. Translated to the context <strong>of</strong> project based<br />
<strong>learning</strong>, we may take these results to indicate that recall <strong>of</strong> project events benefits if<br />
project members can draw on the collaboration expertise developed through their<br />
project <strong>work</strong>. <strong>The</strong> retrospective use <strong>of</strong> collaborative tools well known from day-today<br />
<strong>work</strong> may be part <strong>of</strong> this.<br />
Findings from the research field <strong>of</strong> social contagion <strong>of</strong> memory indicate that there<br />
is a tendency for participants in collaborative remembering to be influenced by others<br />
in the remembering process, individual memory thus being distorted [38, 39]. This<br />
research is however based on experiments far from real-life <strong>work</strong>. Recent research on<br />
memory <strong>of</strong> events that are real, complex <strong>and</strong> significant in people’s lives show that<br />
collective memory is qualitatively different from individual memory in these settings<br />
[40, 41]. In project based <strong>learning</strong>, the inclusion <strong>of</strong> individual as well as collective<br />
steps <strong>of</strong> recall <strong>and</strong> <strong>reflection</strong> may serve as a way <strong>of</strong> exploiting the positive effects <strong>of</strong><br />
recalling alone as well as those <strong>of</strong> doing it in concert.<br />
<strong>The</strong> empirical research on SD student projects outlined in Section 2 indicated that<br />
creating a shared representation based on individual ones results in more knowledge<br />
than that found in the individual, external representations [10]. However, those who<br />
had the most expertise with a certain aspect <strong>of</strong> <strong>work</strong> tended to dominate the team’s<br />
shared, externalized view <strong>of</strong> that aspect <strong>of</strong> <strong>work</strong> [11]. Research stating that expertise<br />
is a combination <strong>of</strong> the expert, his environment <strong>and</strong> the tools he uses [23] underpin<br />
these findings. Apart from being a source <strong>of</strong> power differences, however, expertise is<br />
an important resource for using collaborative tools as cognitive tools for <strong>reflection</strong>.<br />
We conclude from this section that the historical data in collaborative tools are important<br />
resources in unveiling different aspects <strong>of</strong> project <strong>work</strong> involving different<br />
team members, although the value <strong>of</strong> using a particular tool in a specific case must be<br />
considered. A combination <strong>of</strong> individual <strong>and</strong> collective <strong>reflection</strong> seems feasible.<br />
180
426 B.R. Krogstie<br />
4.2 <strong>The</strong> Temporal Distribution <strong>of</strong> Cognition in Retrospective Reflection<br />
Retrospective <strong>reflection</strong> in project based <strong>learning</strong> is part <strong>of</strong> a trajectory <strong>of</strong> project<br />
based <strong>learning</strong> spanning the entire project, but also constitutes a separate sub-activity<br />
within the overall project. <strong>The</strong> distributed cognition <strong>of</strong> retrospective <strong>reflection</strong> can<br />
thus be seen as situated in two different contexts with different objectives for the<br />
ongoing activity: <strong>The</strong>re-<strong>and</strong>-then project <strong>work</strong> focused on completing project tasks,<br />
<strong>and</strong> here-<strong>and</strong>-now retrospective <strong>reflection</strong> focused on making sense <strong>of</strong> there-<strong>and</strong>-then<br />
<strong>work</strong> in the light <strong>of</strong> knowledge <strong>of</strong> the entire process.<br />
Events <strong>of</strong> there-<strong>and</strong>-then <strong>work</strong> may be interpreted as belonging to different subtrajectories,<br />
<strong>and</strong> can be seen as the core <strong>of</strong> what is reflected upon. We illustrate this by<br />
going into some detail about the findings reported in [11] <strong>and</strong> illustrated in Fig.2. In the<br />
retrospective <strong>reflection</strong> <strong>work</strong>shop <strong>of</strong> the team in question, there was a project event not<br />
remembered by any team member in the individual reconstruction <strong>of</strong> the project trajectory<br />
based on memory alone, i.e. not included in any individual timeline. <strong>The</strong> event<br />
marked the onset <strong>of</strong> an activity in the project, more specifically coding <strong>work</strong> done to get<br />
familiar with technology to be used later in the real development. As such, the event<br />
could be seen both as a turning point in the trajectory <strong>of</strong> the entire development activity<br />
<strong>and</strong> as the starting point <strong>of</strong> a sub-trajectory: that <strong>of</strong> the early-investigation <strong>of</strong> technology.<br />
<strong>The</strong> team members’ individual examination <strong>of</strong> historical data in the issue tracking<br />
tool made two <strong>of</strong> them recall the event as important to the project (Fig.2) <strong>and</strong> include<br />
it in their timelines. Later, as the team created a shared project timeline, the event was<br />
included <strong>and</strong> brought up into the discussion <strong>of</strong> how things might have been done<br />
differently in the project. <strong>The</strong> event was discussed in the light <strong>of</strong> the specific activity<br />
for which it marked the starting point <strong>and</strong> in the light <strong>of</strong> the entire development project.<br />
<strong>The</strong> trajectory <strong>of</strong> the project as represented in the shared timeline was compared<br />
to a process described in SD literature as ideal for a certain type <strong>of</strong> development<br />
<strong>work</strong>. <strong>The</strong>se findings illustrate that comparing trajectories is at the core <strong>of</strong> the<br />
retrospective <strong>reflection</strong> process.<br />
Considering retrospective <strong>reflection</strong> as part <strong>of</strong> a trajectory <strong>of</strong> project based <strong>learning</strong><br />
we see a possible effect on tool usage if it is known to the team that historical data<br />
will later on be systematically examined. <strong>The</strong> dual role <strong>of</strong> the tool may lead project<br />
members to adjust their day-to-day tool usage, e.g. by providing more frequent or<br />
elaborate comments associated with ongoing tasks, which may affect project <strong>work</strong> in<br />
a positive way. However, changes may have adverse effects, e.g. if participants cease<br />
to communicate issues in a tool because the logged data may give an unfavourable<br />
impression in retrospect. Adjusting tool use to meet the needs <strong>of</strong> both retrospective<br />
<strong>reflection</strong> <strong>and</strong> day-to-day <strong>work</strong> may be seen as part <strong>of</strong> project based <strong>learning</strong> itself.<br />
We conclude from this section that the ‘return to experience’ <strong>of</strong> retrospective <strong>reflection</strong><br />
involves examination <strong>and</strong> comparison <strong>of</strong> sub-trajectories <strong>of</strong> project <strong>work</strong> <strong>and</strong><br />
that retrospective use <strong>of</strong> tools may affect project <strong>work</strong> through participants’ awareness<br />
<strong>of</strong> the dual role <strong>of</strong> the tools.<br />
4.3 Transformations <strong>of</strong> Representations <strong>and</strong> the Use <strong>of</strong> Cognitive Tools<br />
Looking at the timeline technique as used in [10], <strong>and</strong> illustrated in Fig.1 we see many<br />
examples <strong>of</strong> use <strong>of</strong> representations as cognitive tools. <strong>The</strong> internal, individual<br />
representations <strong>of</strong> the project process were transformed into externalized individual<br />
181
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 427<br />
timelines on paper. Both types <strong>of</strong> representations were used in the collaborative session<br />
through which the knowledge captured in the representations was transformed<br />
into a shared, external representation in the form <strong>of</strong> a timeline, a process likely to<br />
make the individual, internal representations change. <strong>The</strong> internal <strong>and</strong> external representations<br />
created <strong>and</strong> modified through retrospective <strong>reflection</strong> served as cognitive<br />
tools for the teams’ later writing <strong>of</strong> collaborative <strong>reflection</strong> notes. <strong>The</strong> experience<br />
curve helped participants reflect on a particular sub-trajectory corresponding to the<br />
‘attending to feelings’ in the reflective process [31].<br />
<strong>The</strong> above outlined transformations were achieved by use <strong>of</strong> the timeline as a cognitive<br />
tool. <strong>The</strong> successful use <strong>of</strong> timelines in the study indicates that it is a good form<br />
<strong>of</strong> representation to aid both individual <strong>and</strong> collective recall <strong>and</strong> <strong>reflection</strong>.<br />
Salomon [24] argues that “even in the most radical formulations <strong>of</strong> activity-insetting<br />
[..] there is no way to get around the role played by individuals’ representations.”<br />
p.134. <strong>The</strong> timeline approach as conducted in the SD studies pays heed to this<br />
by including the steps <strong>of</strong> creating external, individual representations.<br />
We see the following as favourable characteristics <strong>of</strong> the timeline: the metaphor is<br />
easily grasped by participants, it maps well to the conception <strong>of</strong> process as trajectory,<br />
it is easily understood when presented to others, <strong>and</strong> different timelines representing<br />
the same process are easily combined. <strong>The</strong> timeline provides a frame<strong>work</strong> for contextualizing<br />
historical data in collaborative tools, in which the timing <strong>and</strong> chronological<br />
sequence <strong>of</strong> events associated with the data are generally available.<br />
In the collaborative tools used in project <strong>work</strong>, some <strong>of</strong> the data originating in the<br />
<strong>work</strong> <strong>and</strong> stored in the tools are used as cognitive tools in the <strong>work</strong> process, thus contributing<br />
to the transformation <strong>of</strong> participants’ internal project representations..<br />
Turning to the step <strong>of</strong> retrospective <strong>reflection</strong>, the same data may partially be accessed<br />
in the same way as in day-to-day <strong>work</strong> (as when previous events in the<br />
timeline <strong>of</strong> the issue tracker was examined; this is a tool use similar to day-to-day<br />
coordination <strong>work</strong>) but may also be retrieved in a way not typically used in day-today<br />
<strong>work</strong> (as when early revisions <strong>of</strong> the project wiki main page were chronologically<br />
examined in retrospective <strong>reflection</strong>). Thus, the representations <strong>of</strong> the project<br />
trajectory retrieved from the tool vary with the purpose <strong>and</strong> procedure <strong>of</strong> retrieval.<br />
When collaborative tools are used as cognitive tools for retrospective <strong>reflection</strong>,<br />
externalized representations originally transformed through day-to-day project <strong>work</strong><br />
take part in the transformation <strong>of</strong> other representations <strong>of</strong> the project trajectory;<br />
internal <strong>and</strong> possibly external ones (e.g. timelines), individual <strong>and</strong>/or shared ones.<br />
In the studies outlined in Section 2, tools used to aid retrospective <strong>reflection</strong> were<br />
<strong>of</strong> two types: those created for the purpose <strong>of</strong> retrospective <strong>reflection</strong>, i.e. the timelines<br />
<strong>and</strong> curves on paper <strong>and</strong> whiteboard, <strong>and</strong> tools primarily supporting day-to-day<br />
project <strong>work</strong> <strong>and</strong> being re-used for the purpose <strong>of</strong> retrospective <strong>reflection</strong>, e.g. the<br />
project wiki <strong>and</strong> the issue tracker. In case <strong>of</strong> the latter category <strong>of</strong> tools, even if the<br />
process <strong>of</strong> accessing historical data had some transformative effect on the representations<br />
retrieved, the historical data in the tool remained fixed. <strong>The</strong> timelines <strong>and</strong> curves<br />
created for the purpose <strong>of</strong> retrospective <strong>reflection</strong>, on the other h<strong>and</strong>, underwent<br />
continuous transformation during the retrospective <strong>reflection</strong> activity.<br />
Whereas the empirical studies <strong>of</strong> SD projects indicate that the two types <strong>of</strong> tools<br />
independently benefit retrospective <strong>reflection</strong> [9, 10], there appears to be added value<br />
182
428 B.R. Krogstie<br />
in combining them [11], using the historical data as a means for transforming the<br />
timeline <strong>and</strong> the timeline as a means for making sense <strong>of</strong> the historical data.<br />
By (re-)using a computerized tool that has been used in one context, e.g. that <strong>of</strong><br />
(some aspect <strong>of</strong>) day-to-day project <strong>work</strong>, <strong>and</strong> employing it as a cognitive tool for<br />
retrospective <strong>reflection</strong>, some <strong>of</strong> the expertise inherent in the learner’s there-<strong>and</strong>-then<br />
usage <strong>of</strong> the tool may be applicable to the tool use in retrospective <strong>reflection</strong>. This can<br />
be understood as utilizing the expertise <strong>of</strong> the joint <strong>learning</strong> system [23] <strong>of</strong> day-to-day<br />
<strong>work</strong> in the transition to retrospective <strong>reflection</strong> on that <strong>work</strong>.<br />
<strong>The</strong>re are limits to the insight that can be gained by the use <strong>of</strong> one type <strong>of</strong> representation<br />
<strong>of</strong> a project, e.g. a timeline. This is the case even if the representation is made<br />
more sophisticated (for instance by including representation <strong>of</strong> sub-trajectories like<br />
the individual experience curves). Not all aspects <strong>of</strong> project <strong>work</strong> fit within a temporal/linear<br />
perspective, even if it is possible to revisit most issues along a timeline.<br />
<strong>The</strong>re may be aspects <strong>of</strong> project <strong>work</strong> that are better expressed in other ways, e.g.<br />
with textual descriptions, diagrams or even role play. For instance, a representation<br />
outlining the structure <strong>of</strong> an artefact, showing who contributed to what part, may be<br />
useful to aid <strong>reflection</strong> on the <strong>work</strong> process. Some representations providing synthesized<br />
information about the project are likely to be available in the historical data <strong>of</strong><br />
collaborative tools; project plans are an example. <strong>The</strong> timeline should be seen not<br />
only as a valuable project representation in itself, but also as a good starting point for<br />
identifying <strong>and</strong> creating other representations by helping participants get an overview<br />
<strong>and</strong> create a context for exploration <strong>of</strong> particular issues.<br />
In the studies outlined in Section 2, retrospective <strong>reflection</strong> was conducted when<br />
the projects were more or less finished. Taken from the <strong>learning</strong> objectives <strong>of</strong> the<br />
course, the students were meant to be able to use their experience from the project in<br />
other projects, but implicitly this was taken to happen through individual <strong>learning</strong>.<br />
Revisiting the model <strong>of</strong> the reflective process [31], its outcomes comprise new perspectives<br />
on experiences, change in behaviour, readiness for application, <strong>and</strong> commitment<br />
to action. We would expect a successful process <strong>of</strong> retrospective <strong>reflection</strong> to<br />
result in all <strong>of</strong> these being somehow incorporated in participants’ internal representations<br />
<strong>of</strong> the project experience, <strong>and</strong> some <strong>of</strong> it to be expressed in the external representations<br />
resulting from <strong>reflection</strong>. However, the actual <strong>learning</strong> from experience is<br />
best seen in further project <strong>work</strong>, <strong>and</strong> retrospective <strong>reflection</strong> should be an elements<br />
<strong>of</strong> a <strong>learning</strong> <strong>cycle</strong> [30]. If process improvement is part <strong>of</strong> the purpose <strong>of</strong> retrospective<br />
<strong>reflection</strong>, lessons learned should be captured in representations that are applicable in<br />
later project <strong>work</strong>. How to make lessons learned applicable in practice is a challenging<br />
issue at the core <strong>of</strong> organizational <strong>learning</strong> <strong>and</strong> knowledge management.<br />
Our conclusion from this section is that many types <strong>of</strong> transformations <strong>of</strong> representations<br />
may be involved in, <strong>and</strong> add value to, retrospective <strong>reflection</strong>. <strong>The</strong> transformations<br />
we have discussed were achieved by the use <strong>of</strong> tools primarily supporting<br />
day-to-day <strong>work</strong> as well as tools introduced for retrospective <strong>reflection</strong>.<br />
5 Retrospective Reflection in Project Based Learning: A Model<br />
Based on the previous analysis, we outline a model <strong>of</strong> retrospective <strong>reflection</strong> in<br />
project based <strong>learning</strong> <strong>and</strong> briefly illustrate with examples from Section 2.<br />
183
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 429<br />
<strong>The</strong> model in Fig. 3 illustrates retrospective <strong>reflection</strong> incorporating the main elements<br />
elaborated in Section 4, <strong>and</strong> serves as a generic description <strong>of</strong> retrospective<br />
<strong>reflection</strong> utilizing individual <strong>and</strong> shared timeline representations as well as historical<br />
data in various collaborative tools used in the project. <strong>The</strong> rectangles in the diagram<br />
are representations (internal or external), the ovals are processes in which the representations<br />
are transformed. <strong>The</strong>se processes can be seen as <strong>learning</strong> processes in the<br />
sense that new knowledge is constructed. A representation having an outgoing arrow<br />
pointing to a process is a cognitive tool in that process. Where arrows point in both<br />
directions between a representation <strong>and</strong> a <strong>learning</strong> process, the representation used in<br />
the process is itself changed through the process. <strong>The</strong> finer, dashed arrows show the<br />
use <strong>of</strong> historical data to identify events <strong>and</strong> sub-trajectories.<br />
In more detail, process (1) is the daily project <strong>work</strong>. Collaborative tools are used as<br />
cognitive tools, resulting in data remaining in the tools as historical data. <strong>The</strong> tools in<br />
daily <strong>work</strong> can be seen as continuously updated ‘representations’ <strong>of</strong> the project by<br />
containing data reflecting aspects <strong>of</strong> the project <strong>work</strong>. Internal representations <strong>of</strong> the<br />
project in participants’ heads also impact on, <strong>and</strong> result from, the <strong>work</strong>.<br />
Turning to the retrospective <strong>reflection</strong>, there is an individual step (2) in which individuals<br />
make an external representation <strong>of</strong> the project process in the form <strong>of</strong> a timeline.<br />
We have indicated the possibility to represent sub-trajectories (e.g. a satisfaction<br />
curve). In this, the collaborative tools with their historical data potentially serve as<br />
cognitive tools, as illustrated in Fig.2. In making sense <strong>of</strong> the historical data, the<br />
learner recognizes them as associated with project events <strong>and</strong> sub-trajectories.<br />
Participants’ individual timelines are cognitive tools for collaborative <strong>reflection</strong> (3)<br />
in which a shared timeline is created (Fig. 1), mediated also by the individual, internal<br />
representations. Elements from the individual timelines are included <strong>and</strong> may be<br />
transformed (see Fig.1, the emphasized event). Again, the collaborative tools with<br />
their historical data may be used as cognitive tools. In addition to the shared timeline,<br />
the team may create other representations (e.g. textual <strong>reflection</strong> notes). Some <strong>of</strong> the<br />
resulting representations may serve as cognitive tools in further project <strong>work</strong>, closing<br />
the <strong>cycle</strong> <strong>of</strong> project based <strong>learning</strong>.<br />
<strong>The</strong> model is a simplification with respect to the number <strong>of</strong> representations <strong>and</strong><br />
transformations. For instance, the step <strong>of</strong> orally presenting the contents <strong>of</strong> individual<br />
timelines to the team <strong>and</strong> the transformation <strong>of</strong> historical data that happens through<br />
retrospective use <strong>of</strong> the tool has been only indirectly assumed.<br />
<strong>The</strong> model outlines main elements in the distributed cognition <strong>of</strong> retrospective <strong>reflection</strong><br />
<strong>and</strong> the dynamics among these elements. It can be used to guide the organization<br />
<strong>of</strong> retrospective <strong>reflection</strong> in project-based <strong>learning</strong>. While our analysis addressed<br />
the benefit <strong>of</strong> various representations <strong>and</strong> transformative processes, retrospective<br />
<strong>reflection</strong> may be arranged without the inclusion <strong>of</strong> all the elements. Depending on<br />
constraints <strong>and</strong> objectives, a specific organization <strong>of</strong> the process may leave out the<br />
individual or the collective step <strong>of</strong> creating external timeline representations. Also, the<br />
use <strong>of</strong> collaborative tool history can be omitted. To decide whether a specific collaborative<br />
tool should be used in retrospective <strong>reflection</strong>, a team should take into account<br />
the tool features for data navigation <strong>and</strong> retrieval, tool usage in daily <strong>work</strong>, <strong>and</strong> issues<br />
seen as important to the team <strong>and</strong> thus worth reflecting on.<br />
184
430 B.R. Krogstie<br />
Fig. 3. A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong><br />
6 Conclusion<br />
We have used empirical data from SD student projects to analyse retrospective <strong>reflection</strong><br />
in the context <strong>of</strong> project based <strong>learning</strong> <strong>and</strong> develop a model outlining its<br />
elements <strong>and</strong> dynamics. On basis <strong>of</strong> distributed cognition, the model describes retrospective<br />
<strong>reflection</strong> as a set <strong>of</strong> transformations <strong>of</strong> representations <strong>of</strong> the project process,<br />
internal <strong>and</strong> external. A timeline <strong>of</strong> events is used as a cognitive tool. It fits the idea <strong>of</strong><br />
process trajectories, is good for situating historical data in context, <strong>and</strong> is good for<br />
being shared <strong>and</strong> combined into a collaboratively constructed representation.<br />
<strong>The</strong> scope <strong>of</strong> our analysis <strong>and</strong> model is project based <strong>learning</strong> in which lightweight<br />
collaborative tools are used to support various aspects <strong>of</strong> <strong>work</strong>. While the domain <strong>of</strong><br />
our empirical studies is s<strong>of</strong>tware development, the model may be used to aid the<br />
organization <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> within any domain.<br />
Taking a constructivist view <strong>of</strong> knowledge, there is no single, ‘real’ story <strong>of</strong> a project<br />
that can be revealed through the use <strong>of</strong> timelines <strong>and</strong> collaborative tool history.<br />
Our approach can help participants unveil issues <strong>of</strong> importance to them <strong>and</strong> aid individual<br />
<strong>and</strong> collective sense making <strong>of</strong> various aspects <strong>of</strong> the project. We believe that<br />
the systematic approach, the explicit incorporation <strong>of</strong> individual contributions into<br />
collective, co-constructive <strong>learning</strong> activity, <strong>and</strong> the utilization <strong>of</strong> available resources<br />
in the form <strong>of</strong> historical data together result in <strong>reflection</strong> outcomes that are likely to be<br />
useful with respect to the <strong>learning</strong> objectives <strong>and</strong> valid in the sense that more stones<br />
have been turned.<br />
185
A Model <strong>of</strong> Retrospective Reflection in Project Based Learning 431<br />
<strong>The</strong> analysis presented in this <strong>work</strong> would have benefited from empirical data on<br />
the use <strong>of</strong> more types <strong>of</strong> lightweight collaborative tools in retrospective <strong>reflection</strong> in<br />
SD student projects or similar settings <strong>of</strong> project based <strong>learning</strong>. Further research<br />
should examine the use <strong>of</strong> different collaborative tools within the frame<strong>work</strong> <strong>of</strong> the<br />
model. Results <strong>of</strong> this research can be used to validate the model.<br />
References<br />
1. Blumenfeld, P.C., et al.: Motivating Project-Based Learning: Sustaining the Doing, Supporting<br />
the Learning. Educational Psychologist 26(3 & 4), 369–398 (1991)<br />
2. Boud, D., Keogh, R., Walker, D.: Reflection: Turning Experience into Learning.<br />
Routledge Falmer (1985)<br />
3. Dewey, J.: Democracy <strong>and</strong> education - an introduction to the philosophy <strong>of</strong> education. <strong>The</strong><br />
Free Press, New York (1997) (1916)<br />
4. Schön, D.: Educating the Reflective Practitioner. Jossey-Bass, San Fransisco (1987)<br />
5. Lin, X., et al.: Designing Technology to Support Reflection. Educational Technology, Research<br />
<strong>and</strong> Development 47(3), 43–62 (1999)<br />
6. Edwards, J.S., Shaw, D., Collier, P.M.: Knowledge management systems: finding a way<br />
with technology. Journal <strong>of</strong> Knowledge Management 9(1), 113–125 (2005)<br />
7. Garrett, R.K., Danziger, J.N.: IM=Interruption Management? Instant Messaging <strong>and</strong> Disruption<br />
in the Workplace. Jnl. <strong>of</strong> <strong>Computer</strong>-Mediated Communication 13(1) (2008)<br />
8. Grinter, R.E., Palen, L.: Instant Messaging in Teen Life. In: CSCW 2002, New Orelans,<br />
Louisiana, USA. ACM, New York (2002)<br />
9. Krogstie, B.R.: Using Project Wiki History to Reflect on the Project Process. In: 42nd<br />
Hawaii International Conference on System Sciences, Big Isl<strong>and</strong>, Hawaii. IEEE, Los<br />
Alamitos (2009)<br />
10. Krogstie, B.R., Divitini, M.: Shared timeline <strong>and</strong> individual experience: Supporting retrospective<br />
<strong>reflection</strong> in student s<strong>of</strong>tware engineering teams. In: CSEE&T 2009, Hyderabad (2009)<br />
11. Krogstie, B.R., Divitini, M.: Collaboration tools as a resource for retrospective <strong>reflection</strong><br />
(submitted, 2010)<br />
12. Klein, H.K., Myers, M.M.: A Set <strong>of</strong> Principles for Conducting <strong>and</strong> Evaluating Interpretive<br />
Field Studies in Information Systems. MIS Quarterly 23(1), 67–94 (1999)<br />
13. Derby, E., Larsen, D., Schwaber, K.: Agile Retrospectives. Making Good Teams Great.<br />
Pragmatic Bookshelf (2006)<br />
14. Kasi, V., et al.: <strong>The</strong> post mortem paradox: a Delphi study <strong>of</strong> IT specialist perceptions.<br />
European Journal <strong>of</strong> Information Systems 17, 62–78 (2008)<br />
15. Krogstie, B.R.: <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. In: COOP 2008, Carry-le-<br />
Rouet, Provence, France. Institut d’Etudes Politiques d’Aix-en-Provence (2008)<br />
16. Krogstie, B.: Power through brokering. OSS participation in SE projects. In: ICSE 2008,<br />
Leipzig. IEEE <strong>Computer</strong> Society, Los Alamitos (2008)<br />
17. Krogstie, B., Bygstad, B.: Cross-Community Collaboration <strong>and</strong> Learning in Customer-<br />
Driven S<strong>of</strong>tware Engineering Student Projects. In: CSEE&T 2007, Dublin. IEEE, Los<br />
Alamitos (2007)<br />
18. Halverson, C.: Activity <strong>The</strong>ory <strong>and</strong> Distributed Cognition: Or What Does CSCW Need to<br />
DO with <strong>The</strong>ories? <strong>Computer</strong> Supported Cooperative Work 11, 243–267 (2002)<br />
19. Nardi, B.A.: Studying context: a comparison <strong>of</strong> activity theory, situated action models, <strong>and</strong><br />
distributed cognition. In: St. Petersburg International Workshop on Human-<strong>Computer</strong> Interaction,<br />
St. Petersburg, USSR, Int.Centre Sci. & Tech. Inf, Moscow, Russia (1992)<br />
186
432 B.R. Krogstie<br />
20. Hutchins, E.: Cognition in the Wild. MIT Press, Cambridge (1995)<br />
21. Salomon, G.: Distributed Cognitions. Cambridge University Press, New York (1993)<br />
22. Karasavvidis, I.: Distributed Cognition <strong>and</strong> Educational Practice. Journal <strong>of</strong> Interactive<br />
Learning Research 13(1/2), 11–29 (2002)<br />
23. Kim, B., Reeves, T.C.: Reframing research on <strong>learning</strong> with technology: in search <strong>of</strong> the<br />
meaning <strong>of</strong> cognitive tools. Instructional Science, 35 p., 207–256 (2007)<br />
24. Salomon, G.: No distribution without individuals’ cognition, in Distributed Cognitions. In:<br />
Salomon, G. (ed.) Psychological <strong>and</strong> educational considerations. Cambridge Univ. Press,<br />
Cambridge (1993)<br />
25. Ackerman, M.S., Halverson, C.: Organizational Memory as Objects, Process, <strong>and</strong> Trajectories:<br />
An Examination <strong>of</strong> Organizational Memory in Use. <strong>Computer</strong> Supported Cooperative<br />
Work 13(2), 155–189 (2004)<br />
26. Sharp, H., Robinson, H.: A Distributed Cognition Account <strong>of</strong> Mature XP Teams. In: Abrahamsson,<br />
P., Marchesi, M., Succi, G. (eds.) XP 2006. LNCS, vol. 4044, pp. 1–10.<br />
Springer, Heidelberg (2006)<br />
27. Kirschner, P.A., Erkens, G.: Cognitive tools <strong>and</strong> mindtools for collaborative <strong>learning</strong>.<br />
Journal <strong>of</strong> Educational Computing Research 35(2), 199–209 (2006)<br />
28. Dewey, J.: How we think. A restatement <strong>of</strong> the relation <strong>of</strong> reflective thinking to the educative<br />
process (Revised edn.). D. C. Heath, Boston (1933)<br />
29. Kim, D., Lee, S.: Designing Collaborative Reflection Support Tools in e-project Based<br />
Learning Environment. Journal <strong>of</strong> Interactive Learning Research 13(4), 375–392 (2002)<br />
30. Kolb, D.A., Fry, R.: Towards an applied theory <strong>of</strong> experiental <strong>learning</strong>. In: Cooper, C.L.<br />
(ed.) <strong>The</strong>ories <strong>of</strong> Group Processes, pp. 33–58. John Wiley, London (1975)<br />
31. Boud, D., Keogh, R., Walker, D.: Promoting Reflection in Learning: a Model. In: Boud,<br />
D., Keogh, R., Walker, D. (eds.) Reflection: Turning Experience into Learning, pp. 18–40.<br />
Routledge Falmer (1985)<br />
32. Strauss, A.: Continual permutations <strong>of</strong> action. Aldine de Gruyter, New York (1993)<br />
33. Stahl, G.: Building collaborative knowing. In: Strijbos, J.-W., Kirschner, P.A., Martens,<br />
R.L. (eds.) What We Know About CSCL And Implementing It In Higher Education, pp.<br />
53–85. Kluwer Academic Publishers, Boston (2002)<br />
34. Wegner, D.M.: Transactive memory: A contemporary analysis <strong>of</strong> group mind. In: Mullen,<br />
M.B., Goethals, G.R. (eds.) <strong>The</strong>ories <strong>of</strong> group behaviour, pp. 185–208. Springer, New<br />
York (1987)<br />
35. Hollingshead, A.B.: Retrieval processes in transactive memory systems. Journal <strong>of</strong> Personality<br />
<strong>and</strong> Social Psychology 74, 659–671 (1998)<br />
36. Wegner, D.M., Guiliano, T., Hertel, P.T.: Cognitive interdependence in close relationships.<br />
In: Ickes, W. (ed.) Compatible <strong>and</strong> incompatible relationships. Springer, Heidelberg (1985)<br />
37. Hollingshead, A.B.: Communication, <strong>learning</strong>, <strong>and</strong> retrieval in transactive memory systems.<br />
Journal <strong>of</strong> Experimental Social Psychology 34, 423–442 (1998)<br />
38. Gabbert, F., Memon, A., Allan, K.: Memory conforminty: Can eyewitnesses influence<br />
each other’s memories for an event? Applied Cognitive Psychology 17, 533–543 (2003)<br />
39. Roediger, H.L., Meade, M.L., Bergman, E.T.: Social contagion <strong>of</strong> memory. Psychonomic<br />
Bulletin & Review 8, 365–371 (2001)<br />
40. Alea, N., Bluck, S.: Why are you telling me that? A conceptual model <strong>of</strong> the social function<br />
<strong>of</strong> autobiographical memory. Memory 11, 165–178 (2003)<br />
41. Barnier, A.J., et al.: A conceptual <strong>and</strong> empirical frame<strong>work</strong> for the social distribution <strong>of</strong><br />
cognition: <strong>The</strong> case <strong>of</strong> memory. Cognitive Systems Research 9, 33–51 (2007)<br />
187
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
188
Research paper P8<br />
Title:<br />
Supporting Reflection in S<strong>of</strong>tware Development with Everyday Working Tools<br />
Authors:<br />
Birgit Krogstie <strong>and</strong> Monica Divitini<br />
Published in:<br />
Proceedings <strong>of</strong> COOP 2010<br />
Pages:<br />
20<br />
Complete reference:<br />
Krogstie, B. R. <strong>and</strong> M. Divitini (2010). Supporting Reflection in S<strong>of</strong>tware<br />
Development with Everyday Working Tools. COOP 2010, Aix-en-Provence, France,<br />
19-21 May. Springer.<br />
189
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
190
Supporting Reflection in S<strong>of</strong>tware Development with<br />
Everyday Working Tools<br />
Birgit Krogstie, Monica Divitini<br />
Norwegian University <strong>of</strong> Science <strong>and</strong> Technology,<br />
Sem Sæl<strong>and</strong>s vei 7-9, NO-7491 Trondheim, Norway<br />
{birgitkr, divitini}@idi.ntnu.no<br />
Abstract. Through their day-to-day usage collaboration tools collect data on<br />
the <strong>work</strong> process. <strong>The</strong>se data can be used to aid participants’ retrospective<br />
<strong>reflection</strong> on the process. <strong>The</strong> paper shows how this can be done in s<strong>of</strong>tware<br />
development project <strong>work</strong>. Through a case study we demonstrate how<br />
retrospective <strong>reflection</strong> was conducted by use <strong>of</strong> an industry approach to project<br />
retrospectives combined with the examination <strong>of</strong> historical data in Trac, an<br />
issue tracker. <strong>The</strong> data helped the team reconstruct the project trajectory by<br />
aiding the recall <strong>of</strong> significant events, leading to a shift in the team’s<br />
perspective on the project. <strong>The</strong> success <strong>of</strong> the tool-aided retrospective <strong>reflection</strong><br />
is attributed to its organization as well as the type <strong>of</strong> historical data examined<br />
through the tool <strong>and</strong> the tool features for navigating the data. <strong>The</strong>se insights can<br />
be used to help project teams determine the potential <strong>of</strong> their tools to aid<br />
retrospective <strong>reflection</strong>.<br />
Keywords: <strong>reflection</strong>, project <strong>work</strong>, s<strong>of</strong>tware development, issue tracker.<br />
1 Introduction<br />
S<strong>of</strong>tware development (SD) is highly cooperative <strong>work</strong> which has received<br />
considerable attention in CSCW [1-3]. In SD, mistakes are common as well as costly<br />
[4], <strong>and</strong> <strong>learning</strong> from experience is essential [5].<br />
In SD organizations using large integrated systems to support <strong>and</strong> coordinate <strong>work</strong>,<br />
efforts to have the organization learn from experience may be integrated into<br />
predefined <strong>work</strong> processes <strong>and</strong> supported by tool functionality for collecting <strong>and</strong><br />
providing experience data. <strong>The</strong> idea <strong>of</strong> experience factories [6] reflects a similar idea.<br />
In other settings, as in agile s<strong>of</strong>tware development [7], <strong>learning</strong> from experience may<br />
be more informal <strong>and</strong> deferred to face-to-face project meetings <strong>and</strong> retrospective<br />
<strong>reflection</strong> sessions. Project retrospectives [8-9] in agile SD are an example <strong>of</strong> what is<br />
known in the project management literature as project debriefings [10], which cover<br />
various methods for recording experiences from projects. Project retrospectives are<br />
conducted at the end <strong>of</strong> project phases or when projects are terminated. <strong>The</strong> major<br />
resources are participants’ memory <strong>and</strong> collaborative effort to reconstruct the project,<br />
<strong>and</strong> various facilitation techniques exist (e.g. [11]). In this paper we generically refer<br />
to retrospective <strong>reflection</strong> to indicate activities conducted at the end <strong>of</strong> a project phase<br />
191
to rethink the <strong>work</strong>ing process with the goal <strong>of</strong> <strong>learning</strong> from experience.<br />
Retrospective <strong>reflection</strong> can be seen as an aspect <strong>of</strong> reflective practice [12]<br />
comprising <strong>reflection</strong>-on-action in context <strong>of</strong> the project.<br />
One challenge, found to be among the reasons why many organizations choose not<br />
to conduct retrospective <strong>reflection</strong> [13], is the lack <strong>of</strong> adequate data to help<br />
participants reconstruct the process. Human memory is fallible in recalling a process<br />
<strong>of</strong> weeks or months. For other data sources to be useful in practice, however, they<br />
should provide access to data with an appropriate level <strong>of</strong> detail <strong>and</strong> enough context<br />
to help avoid oversimplification <strong>and</strong> examination <strong>of</strong> unimportant details [13].<br />
<strong>The</strong> objective <strong>of</strong> this paper is to investigate the potential to support project teams’<br />
retrospective <strong>reflection</strong> by the use <strong>of</strong> historical data in lightweight collaboration tools<br />
[14]. <strong>The</strong>se are tools typically providing basic functionality for collaborative activity<br />
without imposing much structure on the process. This allows flexible incorporation <strong>of</strong><br />
the tool into the specific process. Taking lightweight collaboration tools into use<br />
requires little resources (e.g. acquisition cost, time for training) on part <strong>of</strong> users <strong>and</strong><br />
organizations. Lightweight tools are in use in a large number <strong>of</strong> SD projects, some <strong>of</strong><br />
which have very limited resources for organizing retrospective <strong>reflection</strong> sessions.<br />
Lightweight tools may be the choice <strong>of</strong> small organizations, teams doing crossorganizational<br />
<strong>work</strong>, teams basing their selection <strong>of</strong> tools <strong>and</strong> methodologies on local<br />
needs <strong>and</strong> preferences, or teams in organizations running ‘lean’ processes. As a sideeffect<br />
<strong>of</strong> day-to-day <strong>work</strong>, the lightweight collaboration tools collect data originating<br />
in the <strong>work</strong> process. Such historical data may have the appropriate detail <strong>and</strong> context<br />
to aid participants’ retrospective reconstruction <strong>of</strong> the process.<br />
<strong>The</strong> CSCW literature has addressed the use <strong>of</strong> lightweight cooperation technology<br />
to support day-to-day project <strong>work</strong>, e.g. in s<strong>of</strong>tware development [15-17]. Also, there<br />
is research on the use <strong>of</strong> data captured in collaboration tools to gain insights about<br />
collaborative <strong>work</strong> activity, as will be elaborated in the Background section. <strong>The</strong><br />
potential to aid retrospective <strong>reflection</strong> on an entire project process by use <strong>of</strong><br />
historical data in the various lightweight collaboration tools supporting that process<br />
was identified in [18].<br />
For a project team to be able to determine the potential <strong>of</strong> their collaboration tools<br />
to aid retrospective <strong>reflection</strong>, they need knowledge about what it is that makes a tool<br />
useful for such a purpose. Such knowledge is also <strong>of</strong> value to collaboration tool<br />
designers who might want to extend the area <strong>of</strong> application <strong>of</strong> their tools to<br />
retrospective <strong>reflection</strong>.<br />
Research contributions in this area should include empirical investigation <strong>of</strong> tool<br />
use as it unfolds in the context <strong>of</strong> retrospective <strong>reflection</strong>. This paper provides such a<br />
contribution by going into detail on how retrospective <strong>reflection</strong> is aided by the use <strong>of</strong><br />
historical data in a collaboration tool in a SD project followed in an in-depth case<br />
study.<br />
In the background section we outline relevant theory <strong>and</strong> related <strong>work</strong> from the<br />
CSCW field. <strong>The</strong> case section describes the SD project <strong>of</strong> our study <strong>and</strong> the project<br />
management tool used in that project. <strong>The</strong> research method section includes a<br />
description <strong>of</strong> the organization <strong>of</strong> the team’s <strong>reflection</strong> <strong>work</strong>shop. Findings from the<br />
<strong>work</strong>shop are next described, followed by a section on discussion <strong>and</strong> implications,<br />
a section outlining a set <strong>of</strong> guiding questions regarding the potential <strong>of</strong> collaboration<br />
tools to support <strong>reflection</strong> in SD <strong>work</strong>, <strong>and</strong> finally our conclusion.<br />
192
2 Background<br />
In this section we briefly provide a background on distributed cognition as a way <strong>of</strong><br />
underst<strong>and</strong>ing SD <strong>work</strong> <strong>and</strong> on the reconstruction <strong>of</strong> the process trajectory as<br />
important to retrospective <strong>reflection</strong>. We also address how the day-to-day use <strong>of</strong><br />
collaboration technology results in the creation <strong>and</strong> storage <strong>of</strong> historical data.<br />
In the analysis we have used distributed cognition [19-20] to theoretically frame<br />
the problem. Distributed cognition is a perspective which has been used to underst<strong>and</strong><br />
<strong>learning</strong> [21], s<strong>of</strong>tware development [22-23], <strong>and</strong> cooperative <strong>work</strong> more generally<br />
[24]. Cooperative <strong>work</strong> can be seen as distributed across the people involved in the<br />
activity, through coordination <strong>of</strong> the processes between internal <strong>and</strong> external<br />
representations, <strong>and</strong> through time in such a way that results <strong>of</strong> earlier activity can<br />
transform later activity. Individual participants’ mental representations [25] <strong>of</strong> the<br />
project process mediate day-to-day project <strong>work</strong> as well as <strong>reflection</strong> on the <strong>work</strong>.<br />
Strauss [26] argues that when people make sense <strong>of</strong> a process, they think <strong>of</strong> it as a<br />
trajectory. On this basis, retrospective <strong>reflection</strong> can be seen to involve the individual<br />
<strong>and</strong> collective reconstruction <strong>of</strong> the process trajectory, which includes identifying its<br />
significant events. <strong>The</strong> trajectory can have an internal or external representation (e.g.<br />
a diagram or a textual description). In sum, distributed cognition is a particularly<br />
useful frame<strong>work</strong> for analyzing <strong>and</strong> designing for retrospective <strong>reflection</strong> because it<br />
helps shed light on the role <strong>of</strong> representations that mediate the reflective process,<br />
being used <strong>and</strong> transformed by participants individually <strong>and</strong> collaboratively.<br />
In the day-to-day project activity shaping the project trajectory, cooperative <strong>work</strong><br />
entails coordination, access to artifacts in a shared <strong>work</strong>space, <strong>and</strong> formal as well as<br />
informal communication – aspects <strong>of</strong> <strong>work</strong> typically supported by cooperation<br />
technology. Formal communication, for which archiving is important, may for<br />
instance be conducted over email [27] <strong>and</strong> informal communication, e.g. about<br />
ongoing <strong>work</strong>, by synchronous [16, 28] <strong>and</strong> asynchronous chat. A shared <strong>work</strong>space<br />
to hold artifacts under development must be provided, as well as ways <strong>of</strong> providing<br />
co-<strong>work</strong>ers with <strong>work</strong>space awareness [29-31], including social awareness [32].<br />
Coordination <strong>of</strong> complex <strong>work</strong>, e.g. its planning, monitoring <strong>and</strong> re-planning, requires<br />
coordination artifacts [1, 33], computerized tools ranging from simple to-do-lists <strong>and</strong><br />
bulletin boards to advanced enterprise solutions. Among the tools used to aid the<br />
coordination <strong>of</strong> <strong>work</strong> are issue trackers [34-35] used in S<strong>of</strong>tware Development.<br />
As a consequence <strong>of</strong> the activity along the trajectory, data get stored. Depending<br />
on the usage <strong>of</strong> each tool, its data will reflect different aspects <strong>of</strong> the SD <strong>work</strong>, e.g.<br />
development <strong>and</strong> bug-fixing, stakeholder collaboration [36], <strong>and</strong> team-internal,<br />
informal communication [37-38]. From a distributed cognition perspective, the data<br />
are external representations originating in there-<strong>and</strong>-then distributed cognition. <strong>The</strong><br />
data may have the form <strong>of</strong> logs (events, communication), versions <strong>of</strong> project artifacts<br />
(e.g. code, documentation), <strong>and</strong> plans <strong>and</strong> status reports. Data are generally recorded<br />
in repositories that are accessible through the tools. <strong>The</strong> idea <strong>of</strong> capturing information<br />
relevant to a <strong>work</strong> process from tools <strong>and</strong> artifacts already used to support that <strong>work</strong><br />
process has been explored in several ways in the context <strong>of</strong> SD. Some approaches<br />
involve the deliberate tagging <strong>of</strong> information relevant to others [39-40]. <strong>The</strong> use <strong>of</strong><br />
logs from a code management system to support coordination in a SD team has been<br />
proposed [41]: An event notification system <strong>and</strong> a tickertape tool for CVS messages<br />
193
was integrated with the code management system, improving communication in the<br />
team. <strong>The</strong> project memory tool Hipikat is designed to help new members <strong>of</strong> a<br />
distributed SD team learn from more experienced colleagues [42]. <strong>The</strong> memory<br />
contains the artifacts produced during development <strong>and</strong> is built by the tool with little<br />
or no changes to the existing <strong>work</strong> practices. <strong>The</strong> SEURAT system [43] is designed to<br />
capture design rationale in s<strong>of</strong>tware development <strong>and</strong> was originally intended to<br />
support s<strong>of</strong>tware maintenance. <strong>The</strong> system can also trace requirements <strong>and</strong> decisions<br />
[44] in the project. Ar<strong>and</strong>a <strong>and</strong> Venolia [45] found that for purposes <strong>of</strong> coordinating<br />
SD <strong>work</strong>, data repositories from the development process (bug databases <strong>and</strong> email<br />
repositories) could provide some information about the bug fixing ‘stories’, as<br />
reconstructed by the researchers having access to all the repository data, but the data<br />
in the repositories were incomplete <strong>and</strong> sometimes erroneous. <strong>The</strong> study showed that<br />
the reconstruction <strong>of</strong> the stories, including the social, organizational <strong>and</strong> technical<br />
knowledge, required the participants in the bug fixing process to account for the<br />
stories. Ar<strong>and</strong>a <strong>and</strong> Venolia also found that among the most problematic types <strong>of</strong><br />
missing data in the repositories were missing links from the bug records to the source<br />
code involved.<br />
In [18] it was argued that data collected in various collaboration tools in daily use<br />
in a project may be systematically utilized to aid retrospective <strong>reflection</strong> in the project<br />
team. From a distributed cognition perspective, the historical data in the tools can be<br />
considered external representations <strong>of</strong> the project process. Together with team<br />
members’ internal representations <strong>of</strong> the project process <strong>and</strong> external representations<br />
constructed in the team’s collaborative <strong>reflection</strong> effort, the historical data in the tools<br />
are a potential resource in the reconstruction <strong>of</strong>, <strong>and</strong> <strong>reflection</strong> on, the project process.<br />
<strong>The</strong> outcome <strong>of</strong> the <strong>reflection</strong> (i.e. lessons learned, manifest in internal <strong>and</strong> external<br />
representations) subsequently informs the <strong>work</strong> practice, including the use <strong>of</strong> the<br />
collaboration tools from which the historical data were drawn. In the paper, a<br />
lightweight project management tool in the form <strong>of</strong> an issue tracker (Trac [46]) was<br />
used as an example <strong>of</strong> a tool with a potential to support retrospective <strong>reflection</strong>.<br />
<strong>The</strong> aim <strong>of</strong> this paper is to demonstrate in detail how retrospective <strong>reflection</strong> can be<br />
aided by the use <strong>of</strong> historical data in Trac. Outlining how the historical data are<br />
utilized as an integral part <strong>of</strong> a systematic retrospective <strong>reflection</strong> approach with<br />
individual <strong>and</strong> collective steps, we stress how the collaboration tool plays a role in<br />
knowledge building as well as which tool features are particularly useful to support<br />
retrospective <strong>reflection</strong> in the team.<br />
3 Case<br />
Our case is a SD student team in an undergraduate, capstone project course. <strong>The</strong><br />
projects have external customers, require a high level <strong>of</strong> independence in terms <strong>of</strong><br />
choice <strong>of</strong> process, technology <strong>and</strong> solution, <strong>and</strong> are designed to provide a <strong>work</strong> setting<br />
as close to industry SD as possible. Meeting customer requirements is a primary goal,<br />
along with meeting the university’s requirements for certain project deliveries, most<br />
notably a project report to be delivered in a preliminary, mid-term <strong>and</strong> final version.<br />
194
<strong>The</strong>re are 3-5 students in the teams. <strong>The</strong> project takes up half the <strong>work</strong>load in the final<br />
(6 th ) semester <strong>of</strong> a Bachelor <strong>of</strong> IT program.<br />
As part <strong>of</strong> the course, a retrospective <strong>reflection</strong> <strong>work</strong>shop is arranged for each team<br />
at the end <strong>of</strong> the project based on a technique drawn from industry SD practice [8-9].<br />
With this technique, participants individually <strong>and</strong> collectively construct a timeline <strong>of</strong><br />
project events. Opinions <strong>and</strong> feelings about what is positive <strong>and</strong> negative are added,<br />
e.g. by use <strong>of</strong> post-it notes in different colours or by use <strong>of</strong> curves indicating the ‘ups<br />
<strong>and</strong> downs’ <strong>of</strong> the process. <strong>The</strong> final steps <strong>of</strong> the <strong>work</strong>shop are directed at determining<br />
future action based on the analyzed experience. In the <strong>work</strong>shop <strong>of</strong> our course, we use<br />
the following main steps: the participants draw a timeline <strong>of</strong> their project, adding<br />
events considered significant. First, timelines are drawn individually <strong>and</strong> next a<br />
shared one is collectively constructed. <strong>The</strong> emotional aspects <strong>of</strong> the process are<br />
addressed by having participants draw curves (that we call ‘satisfaction curves’) along<br />
the timeline, indicating their satisfaction with the project at different times. Next<br />
follows a discussion <strong>of</strong> issues related to lessons learned. It was shown in a study that<br />
the <strong>work</strong>shops can help the teams share knowledge <strong>and</strong> reach new insights about their<br />
projects [47].<br />
<strong>The</strong> project <strong>of</strong> our study had five team members (Fig. 1). <strong>The</strong>ir task was to develop<br />
an application allowing people to use their mobile phone to keep track <strong>of</strong> the<br />
geographical location <strong>of</strong> their friends <strong>and</strong> engage in textual chat with them. <strong>The</strong><br />
required solution was technically advanced, including a server, an administration<br />
client <strong>and</strong> a mobile client <strong>and</strong> the use <strong>of</strong> services from a provider <strong>of</strong> geographical<br />
positioning data.<br />
Fig. 1: <strong>The</strong> project team <strong>of</strong> the case study: snapshots from everyday project <strong>work</strong><br />
To support the management <strong>of</strong> their development <strong>work</strong>, the team chose to use an<br />
issue tracking tool called Trac (http://trac.edgewall.org/). Trac is a web application<br />
that supports the organization <strong>of</strong> <strong>work</strong> into tasks <strong>and</strong> phases with milestones, <strong>and</strong> is<br />
lightweight in the sense that it contains only the basic features necessary to manage<br />
SD <strong>work</strong> <strong>and</strong> requires little resources to be taken into use. Tasks are defined by tickets<br />
that are assigned to the users, e.g. team members. Trac also contains a wiki, allowing<br />
for any contents to be flexibly added. Trac provides a set <strong>of</strong> views into the files<br />
managed by an underlying file versioning system (SVN) containing e.g. source code<br />
<strong>and</strong> project documents. When a team member uploads (i.e. commits) new or modified<br />
files to the file management system, e.g. from the development environment in which<br />
source code is produced <strong>and</strong> tested, it can be seen in Trac as a new changeset.<br />
Usually, the user adds a comment on what the changeset is about. All versions <strong>of</strong> all<br />
files that have been committed to SVN, as well as the differences between file<br />
versions, can be viewed through Trac.<br />
195
<strong>The</strong> Trac timeline (see Fig. 2, screenshot (a)) contains an item for each changeset<br />
(<strong>and</strong> its comment), wiki update, ticket update, <strong>and</strong> milestone completion. Each item<br />
has a date <strong>and</strong> time <strong>of</strong> the update; an identifier linking to the particular version <strong>of</strong> the<br />
project artifact(s) affected; <strong>and</strong> the name <strong>of</strong> the user doing the update. In Fig. 2 (a), for<br />
example, the comment associated with changeset [86] reflects an update <strong>of</strong> source<br />
code to the SVN server made by Matthew on 13 February. Apart from conveying<br />
status updates, comments are sometimes used more directly for coordination, as in<br />
“be sure to update before doing any changes” (comment made by Justin on 5 May.).<br />
<strong>The</strong> timeline is thus central to the coordination <strong>of</strong> SD <strong>work</strong> in the project, particularly<br />
when team members are not collocated. It provides an instant overview <strong>of</strong> state-<strong>of</strong>affairs<br />
as well as an entry point to the project artifacts.<br />
Fig. 2: Three screenshots from Trac as Justin examines the timeline <strong>and</strong> identifies<br />
pre-study coding<br />
196
4 Research Method<br />
<strong>The</strong> research reported in this paper is interpretive [48]. <strong>The</strong> main source <strong>of</strong> data is a<br />
retrospective <strong>reflection</strong> <strong>work</strong>shop conducted with the selected SD project team in a<br />
usability lab over a period <strong>of</strong> two <strong>and</strong> a half days at the end <strong>of</strong> their project. In<br />
addition, a longitudinal study <strong>of</strong> the team’s <strong>work</strong> had been conducted throughout the<br />
project (one semester), including 45 hours <strong>of</strong> non-participant observation <strong>of</strong> meetings<br />
<strong>and</strong> <strong>work</strong> in the computer lab <strong>and</strong> frequent examination <strong>of</strong> project artifacts (including<br />
the contents <strong>of</strong> Trac) as well as the team’s internal <strong>and</strong> external email communication.<br />
<strong>The</strong> study provided knowledge important to our interpretation <strong>of</strong> data from the<br />
<strong>work</strong>shop. Selection <strong>of</strong> the project was based on a combination <strong>of</strong> the technically<br />
complex project task <strong>and</strong> the team’s decision to use Trac.<br />
Table 1. Organization <strong>of</strong> the <strong>reflection</strong> <strong>work</strong>shop<br />
Session type Step Explanation<br />
Individual 1) Reconstructing<br />
<strong>The</strong> team member was given a sheet <strong>of</strong> paper with a timeline<br />
sessions<br />
the containing a few fixed project events (e.g. delivery deadlines)<br />
(Days 1-2, project <strong>and</strong> was asked to add important events. Next, for each <strong>of</strong> these<br />
each trajectory by events, the researcher added it to a similar timeline on the<br />
session 1,5-<br />
2 hours)<br />
<strong>The</strong><br />
whiteboard<br />
was photographed<br />
memory whiteboard <strong>and</strong> asked the team member to explain the<br />
importance <strong>of</strong> the event. Next, the team member was asked to<br />
draw on the paper sheet a curve showing how he had felt about<br />
<strong>work</strong>ing in the project. Finally he was asked to draw the curve<br />
along the whiteboard timeline, explaining its shape.<br />
2) Using <strong>The</strong> team member was asked to use the laptop <strong>and</strong> Trac to freely<br />
after these collaboration explain the project. When general browsing <strong>and</strong> explanation<br />
sessions tool history to appeared exhaustive, the team member was asked to do a<br />
aid further chronological walkthrough <strong>of</strong> the Trac timeline, exploring links<br />
reconstruction at need. As new events deemed important to the project were<br />
identified, they were added to the whiteboard timeline by the<br />
interviewer, who took care to add events based on what the team<br />
member considered important.<br />
Common 3) Reconstructuring<br />
<strong>The</strong> project timeline was reconstructed again, the researcher<br />
session<br />
the writing down events on the whiteboard as the team members<br />
(Day 3, half<br />
day)<br />
project<br />
trajectory<br />
using<br />
collaboration<br />
tool at need<br />
listed them in a round-robin fashion. A PC with access to Trac<br />
was available to the team members, the screen projected onto the<br />
wall next to the whiteboard. When no more events were<br />
suggested, the team members came to the whiteboard one by one<br />
to draw their satisfaction curve <strong>and</strong> comment on it.<br />
4) Discussion <strong>The</strong>re was a common session addressing questions <strong>of</strong> project<br />
<strong>of</strong> tasks, roles, tasks <strong>and</strong> lessons learned, each question answered by all<br />
lessons team members in turn. Discussion was allowed <strong>and</strong> encouraged.<br />
learned, roles<br />
<strong>The</strong> <strong>work</strong>shop steps were based on the timeline <strong>and</strong> satisfaction curve approach<br />
described in the Case section. Compared to the short retrospective <strong>work</strong>shops<br />
conducted with the other teams in the course, the design incorporated an extra step <strong>of</strong><br />
using historical data in Trac to aid recall. Further, individual timeline construction<br />
was done in a sequence <strong>of</strong> individual sessions (rather than in parallel) to allow the<br />
197
collection <strong>of</strong> rich data about each team member’s timeline construction <strong>and</strong> explore<br />
how the use <strong>of</strong> historical data in Trac might be helpful in different ways or to different<br />
degrees among the team members. <strong>The</strong> lab was equipped like a small meeting room<br />
with a whiteboard <strong>and</strong> a PC. Several movable cameras were installed in the ceiling,<br />
<strong>and</strong> all sessions were videotaped. <strong>The</strong> use <strong>of</strong> Trac was recorded with screen capture<br />
<strong>and</strong> synchronized with the video recording. Photos were taken <strong>of</strong> the whiteboard.<br />
In line with the principle <strong>of</strong> considering multiple interpretations [38], individual<br />
follow-up interviews were made five months after the <strong>work</strong>shop, i.e. when analysis <strong>of</strong><br />
the data had resulted in findings to be presented to the participants <strong>and</strong> discussed in<br />
light <strong>of</strong> their interpretation.<br />
Two questions guided our focus in the <strong>work</strong>shop <strong>and</strong> the subsequent data analysis.<br />
First, we looked for indications that the retrospective use <strong>of</strong> Trac benefited <strong>reflection</strong>,<br />
e.g. helped the team return to experience, re-evaluate it <strong>and</strong> create new perspectives.<br />
More specifically, we looked for events recalled only by some team members, events<br />
recalled only after use <strong>of</strong> Trac <strong>and</strong>, among those, events <strong>of</strong> general importance in<br />
later discussion, e.g. addressing lessons learned. Second, we wanted to know more<br />
about how Trac was used by the team members to aid their recall <strong>and</strong> <strong>reflection</strong>, i.e.<br />
what particular functionality <strong>and</strong> data in the tool seemed relevant <strong>and</strong> useful. In the<br />
data analysis we systematically examined the individual timelines <strong>and</strong> the shared<br />
timeline <strong>and</strong> the recordings <strong>of</strong> the sessions in which they were made, making a table<br />
<strong>of</strong> all events that were mentioned, noting who had mentioned them <strong>and</strong> whether they<br />
were mentioned before or after examination <strong>of</strong> the data in Trac. We picked one event,<br />
namely pre-study coding, based on the following criteria: the event was recalled only<br />
after use <strong>of</strong> Trac, it was recalled by some team members only, <strong>and</strong> it was important in<br />
the team’s final discussion <strong>of</strong> lessons learned, <strong>The</strong> selection <strong>of</strong> an event guided further<br />
analysis, including close examination <strong>of</strong> all <strong>work</strong>shop data relating to the event with<br />
the aim <strong>of</strong> to providing a coherent account <strong>of</strong> how the event was elaborated in, <strong>and</strong><br />
impacting on, the <strong>reflection</strong> process in the team, Data were transcribed <strong>and</strong> translated<br />
into English at need. In the presentation, names (including user names in screenshots)<br />
are made anonymous.<br />
Our study can be seen as a “fine-grained, field-based empirical study” useful for<br />
designing systems <strong>and</strong> underst<strong>and</strong>ing the nuances <strong>of</strong> practice [24]. Even though the<br />
<strong>reflection</strong> <strong>work</strong>shop in our case was organized with a clear research agenda,<br />
impacting e.g. on its duration, the participants did real <strong>reflection</strong> on their SD project.<br />
<strong>The</strong>re are two main limitations to the study. First, it was conducted in the context<br />
<strong>of</strong> SD education. <strong>The</strong> project was however close to industry practice, e.g. with a<br />
customer <strong>and</strong> requirements for a <strong>work</strong>ing s<strong>of</strong>tware product, <strong>and</strong> the team had some<br />
members who were relatively experienced as developers. Second, a single case study<br />
limits the possibilities to generalize from results. However, our purpose was to<br />
underst<strong>and</strong> the details <strong>of</strong> a team’s <strong>reflection</strong> on their project <strong>and</strong> investigate the<br />
potential to support <strong>reflection</strong> through certain steps <strong>and</strong> with certain resources<br />
198
5 Findings<br />
In this section we draw on our <strong>work</strong>shop data <strong>and</strong> describe the project team’s recall <strong>of</strong><br />
<strong>and</strong> <strong>reflection</strong> on ‘pre-study coding’, an example <strong>of</strong> a project event recalled only<br />
through the aid <strong>of</strong> historical data in the collaboration tool. To provide some context,<br />
we start by describing the team’s project process on basis <strong>of</strong> the longitudinal study.<br />
5.1 <strong>The</strong> Process <strong>and</strong> Organization <strong>of</strong> the Case Project<br />
<strong>The</strong> members were generally satisfied with their team <strong>and</strong> the assigned task. <strong>The</strong><br />
relationship with supervisor <strong>and</strong> customer <strong>work</strong>ed well throughout the project.<br />
Requirements for the product were more or less clear after the second customer<br />
meeting. Prior to midterm report delivery, time in the project was spent partially on<br />
trying out technology, later to be characterized by the team as ‘pre-study coding’,<br />
partially on writing the project report. After four iterations, the team delivered a<br />
product with which both they <strong>and</strong> their customer were satisfied.<br />
<strong>The</strong> team had decided on a flat organization, formally with Eric as project manager.<br />
In practice, Justin had that role, taking main responsibility e.g. for deliveries <strong>and</strong><br />
meeting minutes. Matthew was seen as the technical expert, taking responsibility for<br />
the server application <strong>and</strong> the overall design. Justin was in charge <strong>of</strong> the client<br />
application, <strong>and</strong> Travis for the map functionality implemented late in the project. All<br />
five team members actively participated in programming <strong>and</strong> report writing. Fig. 3<br />
(created by the authors processing information from the Trac timeline) shows that all<br />
team members participated <strong>and</strong> that Matthew <strong>and</strong> Justin did the most ticket updates.<br />
Fig. 3: A summary overview <strong>of</strong> the Trac timeline contents<br />
5.2 Pre-study Coding in the Individual, Memory-based Accounts<br />
<strong>The</strong> five individual whiteboard timelines created on the basis <strong>of</strong> memory alone<br />
(<strong>work</strong>shop Step 1, see Table 1) vary in terms <strong>of</strong> the selection <strong>of</strong> events <strong>and</strong> the shape<br />
<strong>of</strong> the satisfaction curves. <strong>The</strong> events remembered to some extent reflect the roles in<br />
the project. For instance, Matthew is the only one who remembers choices <strong>of</strong> certain<br />
server technology. In comparison, a cancelled feature freeze that had a high impact on<br />
199
the entire team, is remembered by four out <strong>of</strong> five members. <strong>The</strong> onset <strong>of</strong> pre-study<br />
coding is not mentioned as important by anyone at this point. <strong>The</strong> onset <strong>of</strong> ‘coding’ is<br />
however mentioned by Matthew <strong>and</strong> Justin. Matthew refers to a couple <strong>of</strong> events<br />
involving the choice <strong>of</strong> technology in the project before midterm, but says nothing<br />
explicitly about coding in this period. He places ‘coding startup’ right after midterm.<br />
Justin places ‘code start’ on his timeline just before midterm. “Before that we had<br />
only been <strong>work</strong>ing with report, <strong>and</strong> then we actually started writing code.” When<br />
accounting for the impact <strong>of</strong> a necessary change <strong>of</strong> frame<strong>work</strong> some time after<br />
midterm, Justin explains that most <strong>of</strong> the coding <strong>of</strong> the final product took place during<br />
a short period towards the end <strong>of</strong> the project.<br />
5.3 Pre-study Coding Emerging through Examination <strong>of</strong> Trac<br />
With the help <strong>of</strong> Trac in Step 2, all team members identify new events. In addition,<br />
there are some modifications <strong>of</strong> names <strong>of</strong> existing ones <strong>and</strong> changes <strong>of</strong> their timeline<br />
position. Justin <strong>and</strong> Matthew are the team members for which the examination <strong>of</strong> the<br />
timeline brings up most new events. As they chronologically examine the timeline,<br />
Justin <strong>and</strong> Matthew both examine the changesets [81] <strong>and</strong> [82] made by Matthew on<br />
13 February. <strong>The</strong> comment ‘Initial import’ associated with both changesets indicates<br />
that they are about the inclusion <strong>of</strong> files needed to start producing <strong>and</strong> testing Java<br />
code. Fig. 2 shows the sequence <strong>of</strong> screenshots as Justin investigates the timeline item<br />
<strong>of</strong> changeset [82]. Starting from the timeline view (a), he clicks [82]. <strong>The</strong> window (b)<br />
now displays what has been changed; a list <strong>of</strong> files. As a programmer Justin knows<br />
that the coding done by Matthew is in FindPerAnton.java, the rest <strong>of</strong> the files being<br />
files necessary to set up a running program <strong>and</strong> included for the first time in the initial<br />
import. He clicks the filename, <strong>and</strong> the window changes to show the file contents (c).<br />
Examining the code, Justin exclaims: “Oh yes, it was only, kind <strong>of</strong> ‘Hello world’ on<br />
the mobile” <strong>and</strong> returns to the timeline (a). ‘Pre-study coding’ is then added to the<br />
whiteboard timeline (see Fig. 4).<br />
Fig. 4. Justin’s <strong>and</strong> Matthew’s timelines. Frames/emphasis <strong>of</strong> arrows show<br />
additions/modifications after examination <strong>of</strong> Trac.<br />
Matthew, on examining the same item during his timeline construction, says:<br />
“Right there, we have started coding. It was probably pre-study coding.” He suggests<br />
that the event be placed in the first half <strong>of</strong> the pre-midterm project period (see Fig. 4).<br />
200
As a consequence, ‘startup coding’, already on Matthew’s timeline just after midterm,<br />
is changed to ‘startup coding according to requirements spec (start iterations)’.<br />
5.4 Pre-study Coding in the Co-constructed Account<br />
In the collective session <strong>of</strong> Step 3, most events from the individual timelines were<br />
included in the shared timeline. Occasionally, the team on their own initiative looked<br />
into Trac to check details, typically the timing <strong>of</strong> events. <strong>The</strong> issue <strong>of</strong> pre-study<br />
coding was explicitly addressed. Justin early on repeated his statement that coding did<br />
not start until midterm, <strong>and</strong> this was not questioned by anyone. This can be<br />
interpreted as a shared underst<strong>and</strong>ing in the team that ‘coding’ (as a st<strong>and</strong>alone term)<br />
meant ‘coding in accordance with requirements specification’. As part <strong>of</strong> the roundrobin<br />
provision <strong>of</strong> events, ‘pre-study coding’ was mentioned by Justin <strong>and</strong> added to<br />
the timeline. Eric suggested another name for the same event: “Un<strong>of</strong>ficial coding”.<br />
This received no further comments. In the later step <strong>of</strong> drawing satisfaction curves<br />
along the whiteboard, Justin explicitly referred to pre-study coding as he made an<br />
upward curve bend early in the project, explaining that he enjoyed the pre-study<br />
coding because they got things <strong>work</strong>ing then.<br />
5.4 Pre-study Coding in Reflections on Lessons Learned<br />
<strong>The</strong> topic <strong>of</strong> pre-study coding was recurrent in the discussion in Step 4, particularly<br />
when the team addressed lessons learned. Justin suggested that they might have<br />
overrated their own capabilities by starting coding as late as midterm. Matthew said<br />
that more pre-study coding might have helped the team avoid the extra <strong>work</strong> <strong>of</strong><br />
<strong>learning</strong> a new frame<strong>work</strong>, because they would have known better the limitations <strong>of</strong><br />
the technology. Overall in the discussion, pre-study coding was given a central role in<br />
the account <strong>of</strong> what happened in the project <strong>and</strong> in conceptions <strong>of</strong> how to ideally run a<br />
similar project. <strong>The</strong> considerations on pre-study coding were made by Justin <strong>and</strong><br />
Matthew. <strong>The</strong>re seemed to be full agreement in the team that pre-study coding is<br />
necessary to reach the technical competence needed for producing design <strong>and</strong> code<br />
when a development task requires use <strong>of</strong> technology new to the team.<br />
5.5 Insights from the Follow-up Interviews<br />
In the follow-up interviews with the team members, they were presented with the<br />
researcher’s interpretation that the <strong>work</strong>shop had lead to a shift in the team’s attention<br />
to pre-study coding <strong>and</strong> conception <strong>of</strong> its significance to the project. All team<br />
members expressed that this finding is in line with their own interpretation <strong>of</strong> what<br />
happened in the <strong>work</strong>shop. <strong>The</strong> interviews with Eric, Travis <strong>and</strong> Kyle however<br />
revealed some viewpoints that had not been brought into the team’s discussion. When<br />
Eric during the collective reconstruction <strong>of</strong> the timeline had used the term ‘un<strong>of</strong>ficial<br />
coding’ about the pre-study coding, he did not refer to the fact that the team deemphasized<br />
this activity but to the fact that pre-study coding had been conducted by<br />
201
some team members only. In Eric’s opinion, the project would have gained from a<br />
more distributed effort, making the entire team more competent with the technology<br />
at an earlier point. Travis <strong>and</strong> Kyle expressed similar views. While these viewpoints<br />
might have matured in the period after the <strong>reflection</strong> <strong>work</strong>shop, the <strong>work</strong>shop<br />
recordings indicate that the views were there during the <strong>work</strong>shop but had never<br />
properly entered the discussion.<br />
6 Discussion <strong>and</strong> Implications<br />
<strong>The</strong> example in the Findings illustrates how collaboration technology supporting dayto-day<br />
<strong>work</strong> practice in a SD project contained data that could be utilized by team<br />
members in a systematic <strong>and</strong> collaborative effort to reconstruct <strong>and</strong> reflect upon the<br />
project process.<br />
<strong>The</strong> event called ‘pre-study coding’ was identified only through the aid <strong>of</strong><br />
collaboration tool history <strong>and</strong> only by two <strong>of</strong> the five team members during the<br />
individual timeline construction. <strong>The</strong> event was later brought into the team’s shared<br />
timeline, was referred to by a team member as impacting on his personal experience<br />
<strong>of</strong> the project, <strong>and</strong> was central in the team’s <strong>reflection</strong>s on lessons learned. Comparing<br />
the team members’ original accounts <strong>of</strong> the project at the outset <strong>of</strong> the <strong>work</strong>shop with<br />
the final discussion, there was a clear difference: the significance <strong>of</strong> pre-study coding,<br />
in this project in particular <strong>and</strong> in SD projects in general, had become clearer. We see<br />
this as an example <strong>of</strong> knowledge relevant to the SD practice being created through<br />
retrospective <strong>reflection</strong>.<br />
In this section we address why the tool-aided retrospective <strong>reflection</strong> was<br />
successful. We do this by considering the <strong>work</strong> practice, including the retrospective<br />
<strong>reflection</strong> effort itself, as a case <strong>of</strong> distributed cognition [18], <strong>and</strong> by looking into the<br />
type <strong>of</strong> historical data available in the tool, the tool features used to retrieve the data<br />
in the <strong>reflection</strong> <strong>work</strong>shop, <strong>and</strong> the organization <strong>of</strong> the <strong>work</strong>shop.<br />
6.1 Historical Data in the Collaboration Tool<br />
Cognition involved in a <strong>work</strong> practice is distributed over participants <strong>and</strong> the artifacts<br />
involved. Accordingly, it will typically be necessary to examine different data sources<br />
to make sense <strong>of</strong> an issue or event associated with that practice. In our case, what<br />
made Justin recognize that coding had happened in the project, was an item in the<br />
Trac timeline (see Fig. 2). <strong>The</strong> type <strong>of</strong> timeline item (a change to a file in the<br />
underlying file versioning system) <strong>and</strong> the comment following the change (i.e. the<br />
comment explicitly added by the user making the change, for purposes <strong>of</strong> day-to-day<br />
coordination) made it possible for him to quickly infer what this was about. Following<br />
the link, he recognized the element <strong>of</strong> interest in the list <strong>of</strong> files, <strong>and</strong> selecting the file<br />
in its there-<strong>and</strong>-then version, he could see what the piece <strong>of</strong> code was actually doing.<br />
<strong>The</strong> combination <strong>of</strong> the data <strong>and</strong> Justin’s background (as a programmer, Trac user,<br />
<strong>and</strong> co-creator <strong>of</strong> the data in front <strong>of</strong> him), enabled him to reconstruct the state <strong>of</strong> the<br />
project <strong>and</strong> identify the event.<br />
202
<strong>The</strong> type <strong>of</strong> data involved included 1) data used for day-to-day coordination <strong>of</strong><br />
project <strong>work</strong>, reflecting ongoing tasks (e.g timeline items with comments, made by<br />
different team members; the list <strong>of</strong> artifacts to which the selected timeline update<br />
applies) 2) data describing the state <strong>of</strong> an artifact in the project <strong>work</strong>space (e.g. the<br />
there-<strong>and</strong>-then version <strong>of</strong> the source code file). Different aspects <strong>of</strong> the development<br />
<strong>work</strong> (e.g. project management, coding) being distributed in accordance with team<br />
roles, the data used to identify pre-study coding reflects a social distribution <strong>of</strong> the<br />
cognitive processes involved in the onset <strong>of</strong> pre-study coding.<br />
<strong>The</strong> combination <strong>of</strong> data originating in coordination <strong>and</strong> data from the<br />
development <strong>work</strong>space is useful for identifying events associated with development.<br />
A type <strong>of</strong> data less relevant for the identification <strong>of</strong> pre-study coding, but that<br />
frequently evoked team members’ recall <strong>and</strong> reaction in the <strong>work</strong>shop <strong>of</strong> our study,<br />
are timeline comments addressing the social/emotional side <strong>of</strong> <strong>work</strong> <strong>and</strong> thus<br />
providing social awareness in day-to-day <strong>work</strong>.<br />
Comparing the data examined in the <strong>work</strong>shop sequence illustrated in Fig. 2 with<br />
the aggregate data on the project process shown in Fig. 3, we note that the aggregate<br />
data could not have been used to provide context for single events. <strong>The</strong> diagram,<br />
though a useful resource, provides only a partial overview <strong>of</strong> the process. In our case,<br />
the aggregate diagram shows that all team members participated. What cannot be seen<br />
from the diagram is the exact type <strong>and</strong> amount <strong>of</strong> <strong>work</strong> done by each team member.<br />
For instance, we knew from observation <strong>of</strong> the team that Matthew tended to make<br />
updates in large increments, uploading new code to the shared server after <strong>work</strong>ing on<br />
it locally on his own machine for long, sometimes for days. While the diagram thus<br />
seems to indicate that Matthew <strong>and</strong> Eric did a similar amount <strong>of</strong> coding in the project,<br />
Matthew in reality did more coding than Eric. Another example is that when two team<br />
members were <strong>work</strong>ing in pairs, one <strong>of</strong> them tended to be in charge <strong>of</strong> the keyboard.<br />
In sum, the aggregate timeline data provides an additional overview. To get<br />
adequately detailed <strong>and</strong> contextualized historical data from Trac for the recall <strong>of</strong><br />
specific project events, however, an examination <strong>of</strong> individual timeline items is<br />
necessary. It should also be noticed that the fact that project <strong>work</strong> was sometimes<br />
conducted in pairs is not reflected anywhere in Trac. <strong>The</strong> day-to-day usage <strong>of</strong> the tool<br />
provides it with potential to be used in shedding light on some aspects <strong>of</strong> the project<br />
process - but not all.<br />
6.2 Navigating the Historical Data in the Collaboration Tool<br />
It was through chronological traversal <strong>of</strong> the timeline events that Justin identified the<br />
first uploading <strong>of</strong> source code to the shared <strong>work</strong>space. Also, the timeline provided a<br />
condensed overview, including the possibility – unexploited in this case - to filter the<br />
type <strong>of</strong> events displayed. From the timeline item <strong>of</strong> interest, it took just a click to get<br />
more detailed, but still overview level information about the change, e.g. which files<br />
had been changed. <strong>The</strong> usefulness <strong>of</strong> this feature fits with the point made in [45] that<br />
missing links from bug records to source code is highly problematic for the<br />
reconstruction <strong>of</strong> the bug fixing process. Going up <strong>and</strong> down between different levels<br />
<strong>of</strong> detail thus proved important. From the overview, there was direct access to the<br />
state <strong>of</strong> the artifact (e.g. the selected file) at the time <strong>of</strong> the change. Switching between<br />
203
the modes <strong>of</strong> traversal, overview, <strong>and</strong> exploring the state <strong>of</strong> the artifact was easily<br />
achieved. <strong>The</strong> ease with which several months’ <strong>work</strong> can thus be navigated in Trac by<br />
participants contrasts with the effort needed to examine less structured data on<br />
collaborative <strong>work</strong>, e.g. video recorded meetings [49].<br />
While the data in the tool reflects socially distributed cognition, the timeline<br />
reflects the temporal distribution <strong>of</strong> cognition <strong>and</strong> is a starting point for identifying<br />
significant events. Another aspect <strong>of</strong> the temporarily distributed cognition is that<br />
when the tool is used retrospectively to make sense <strong>of</strong> data, familiarity with the tool<br />
provides context for the user. Justin had been using the Trac timeline daily throughout<br />
the project to get an overview <strong>of</strong> current project status. Justin’s experience <strong>and</strong> skills<br />
in how to make sense <strong>of</strong> the process through the collaboration tool mediated his<br />
retrospective <strong>reflection</strong>.<br />
6.3 Appropriately Organizing the Process <strong>of</strong> Retrospective Reflection<br />
Essential to the success <strong>of</strong> the <strong>reflection</strong> <strong>work</strong>shop was its organization. <strong>The</strong> example<br />
in Fig. 2 shows individual use <strong>of</strong> Trac to reconstruct a timeline, but as a whole, our<br />
study shows how historical data successfully plays a role in supporting cooperative<br />
<strong>work</strong>. A core idea in the approach [8-9] on which the <strong>work</strong>shop was based is that<br />
individual perspectives should be utilized in the construction <strong>of</strong> a shared<br />
underst<strong>and</strong>ing. In the study, individual timelines mediated individual recall <strong>and</strong><br />
<strong>reflection</strong> as well as collective timeline construction.<br />
From a distributed cognition perspective, the individual <strong>and</strong> shared timelines can<br />
be seen as external representations <strong>of</strong> the project trajectory, <strong>and</strong> the steps <strong>of</strong> the<br />
<strong>work</strong>shop as a process <strong>of</strong> coordinating the cognitive processing between internal <strong>and</strong><br />
external representational states. In our study, we looked at a collaboration tool as a<br />
resource for the cognitive processing between internal <strong>and</strong> external representational<br />
states. This happens in three ways, as illustrated in Fig. 5: <strong>The</strong> day-to-day <strong>work</strong>, seen<br />
as a process trajectory along which there are multiple points <strong>of</strong> action <strong>and</strong> interaction,<br />
leaves a trace <strong>of</strong> external representations <strong>of</strong> the distributed cognition involved, in the<br />
form <strong>of</strong> historical data in the tool. This is a result <strong>of</strong> Trac’s role in supporting the dayto-day<br />
cooperative project <strong>work</strong> (1). In the process <strong>of</strong> retrospective <strong>reflection</strong>, Trac is<br />
used as an aid to individuals’ recall <strong>of</strong> project events (2), resulting in the creation <strong>of</strong><br />
the individual timelines, which are also external representations <strong>of</strong> the project<br />
trajectory. <strong>The</strong> individual timelines are next used as a resource when the team<br />
collaboratively creates another external representation <strong>of</strong> the project trajectory: the<br />
shared timeline. To aid the alignment <strong>of</strong> the individual timelines, Trac is used<br />
collaboratively by the team as an aid to clarify details in the construction <strong>of</strong> a shared<br />
representation <strong>of</strong> the project (3), e.g. exact dates <strong>and</strong> times <strong>of</strong> specific events. Fig. 5 is<br />
an instantiation <strong>of</strong> the more generic model <strong>of</strong> retrospective <strong>reflection</strong> outlined in [18].<br />
Used in retrospective <strong>reflection</strong> on day-to-day development <strong>work</strong> the collaboration<br />
tool serves as a boundary object between that practice <strong>and</strong> retrospective <strong>reflection</strong><br />
[50]. Developers’ awareness that the tool will be used retrospectively may affect their<br />
day-to-day <strong>work</strong> practice [24]. For instance, s<strong>of</strong>tware developers may start providing<br />
more informative comments about their updates to make it easier to underst<strong>and</strong> them<br />
retrospectively. This is likely to benefit day-to-day coordination. A change with<br />
204
potentially negative effects is if developers, to make a good retrospective impression<br />
<strong>of</strong> their <strong>work</strong>, start hiding information. For instance, mistakes in intermediate stages<br />
<strong>of</strong> <strong>work</strong> can be hidden by making less frequent updates or by avoiding commenting<br />
about error corrections (a type <strong>of</strong> comment common in the Trac timeline in our<br />
study). This could adversely affect day-to-day <strong>work</strong>space awareness as well as the<br />
possibility to get a realistic picture <strong>of</strong> the development process from the Trac timeline.<br />
Fig. 5: Trac in a dual role as a collaboration tool in a project: supporting day-to-day<br />
<strong>work</strong> <strong>and</strong> retrospective <strong>reflection</strong> on that <strong>work</strong>. Adapted from [18].<br />
Our findings indicated that the team in our study through the tool-aided<br />
retrospective <strong>reflection</strong> succeeded in creating new knowledge about their s<strong>of</strong>tware<br />
development practice. However, the individual differences in respect <strong>of</strong> how helpful<br />
Trac was in the recall <strong>of</strong> events, point to challenges <strong>of</strong> uneven power distribution in<br />
the team. <strong>The</strong> benefit <strong>of</strong> using a development-oriented collaboration tool to aid<br />
<strong>reflection</strong> appeared greater for those dominating the development <strong>work</strong>. Strong<br />
involvement in the activity from which the data originated provided them with<br />
context for recognizing the data as connected to project events. Having identified the<br />
pre-study coding event, the two lead programmers seemed to retain ownership <strong>of</strong> the<br />
event in the team’s discussion, which can explain why alternative viewpoints on prestudy<br />
coding among some team members were never properly disclosed.<br />
205
7 <strong>The</strong> Potential <strong>of</strong> Collaboration Tools to Support Reflection in SD<br />
Work: A Set <strong>of</strong> Guiding Questions<br />
<strong>The</strong> previous sections outlined how certain tool features as well as the type <strong>of</strong> data<br />
stored in the tools <strong>and</strong> the organization <strong>of</strong> the <strong>reflection</strong> activity were important when<br />
historical data in Trac was used to aid the reconstruction <strong>of</strong> the project trajectory in<br />
retrospective <strong>reflection</strong>. In this chapter we ask: in what way are these findings useful<br />
from a perspective <strong>of</strong> supporting <strong>reflection</strong> on SD <strong>work</strong> more generally?<br />
Our answer is tw<strong>of</strong>old. First, we believe that the results hold promise for the use <strong>of</strong><br />
tools like Trac to assume a dual role <strong>of</strong> supporting daily <strong>work</strong> as well as retrospective<br />
<strong>reflection</strong> in SD projects. It was relatively unproblematic in our study to include steps<br />
<strong>of</strong> investigating collaboration tool history into an existing approach to organizing<br />
retrospective <strong>reflection</strong>. To make the tool-aided version <strong>of</strong> the approach integral to SD<br />
practice, the <strong>work</strong>shop organization must be adjusted to feasible schedule, i.e. one<br />
that can be followed under resource constraints in real projects.<br />
Second, we believe that the findings can be generalized to the use <strong>of</strong> collaboration<br />
tools in retrospective <strong>reflection</strong> in SD <strong>work</strong>, answering the intent to provide<br />
empirically grounded insights to project teams <strong>and</strong> tool designers about what makes<br />
collaboration tools useful for such a purpose. In Section 6, we structured the findings<br />
into some issues that seem crucial to the success <strong>of</strong> the selected collaborative tool to<br />
aid retrospective <strong>reflection</strong>. Based on this underst<strong>and</strong>ing, we outline questions that<br />
point to essential aspects <strong>of</strong> how a collaboration tool <strong>and</strong> its historical data can<br />
support retrospective <strong>reflection</strong>. <strong>The</strong> objective is not to find the ‘best’ tool for the<br />
support <strong>of</strong> <strong>reflection</strong> but to help users <strong>and</strong> designers identify, or modify, strengths <strong>and</strong><br />
weaknesses <strong>of</strong> the tools for such use. To illustrate how the answers may differ for<br />
various lightweight collaborative tools, we refer to the research literature on some<br />
tools <strong>of</strong>ten used in project <strong>work</strong>: project wikis, instant messaging, <strong>and</strong> email.<br />
What aspects <strong>of</strong> the project process are reflected in the historical data in the<br />
tool? Which SD challenges worth reflecting upon have been addressed by use <strong>of</strong> this<br />
tool? An issue tracker contains data that may be used to aid reconstruction <strong>of</strong> the<br />
development process, with a focus on its technical <strong>and</strong> formal aspects. <strong>The</strong> informal<br />
communication is likely to be reflected in other tools, e.g. instant messaging tools [28,<br />
51]. Other lightweight project management tools, e.g. project wikis [52], might reflect<br />
more <strong>of</strong> the informal collaboration in the team than what is documented in the<br />
historical data <strong>of</strong> an issue tracker. Stakeholder collaboration, a frequently challenging<br />
SD issue [36], may partially be reflected in email logs.<br />
Does the tool provide features for getting a chronological overview <strong>of</strong> aspects<br />
<strong>of</strong> the project process? As seen in our study, the traversal <strong>of</strong> a chronological<br />
overview can be used to structure the process <strong>of</strong> examining historical data.<br />
Chronological overviews are essential in project management <strong>and</strong> thus provided by<br />
project management tools such as Trac. A project wiki, by being a wiki, contains<br />
functionality for getting an overview <strong>of</strong> the revisions <strong>of</strong> each page [52]. <strong>The</strong><br />
disadvantage <strong>of</strong> browsing through numerous revisions <strong>of</strong> a wiki page is that changes<br />
may be small <strong>and</strong> uninteresting. Conversely, increments reflecting the process as it<br />
206
unfolded may shed light on the process in a way complementing more formal<br />
overviews. An email application allows for the filtering <strong>of</strong> email messages to be<br />
displayed in a mailbox, enabling for instance rapid overview <strong>of</strong> all email sent to the<br />
project customer or all messages with a particular keyword in its title (e.g. ‘status<br />
report’). IM tools generally provide poor overview information, but the log contents<br />
are chronological <strong>and</strong> time-stamped.<br />
Does the tool provide features for accessing the project artifacts in their there<strong>and</strong>-then<br />
versions? One <strong>of</strong> the strengths <strong>of</strong> Trac as a SD tool is the way task<br />
organization links to project artifacts, stressing the close connection between process<br />
<strong>and</strong> product in SD <strong>work</strong>, the complexity <strong>of</strong> the process depending on the complexity<br />
<strong>of</strong> the product [1]. A project wiki may provide direct access to some project artifacts<br />
[17], but is not likely to provide read access to the entire development-related<br />
<strong>work</strong>space with its artifacts in their there-<strong>and</strong>-then versions, as does Trac. Email<br />
frequently includes relevant project artifacts as attachments. Instant messaging logs<br />
might contain elements <strong>of</strong> project artifacts (e.g. source code excerpts) [51].<br />
Does the tool provide features for easy navigation between overview <strong>and</strong><br />
detail? <strong>The</strong> possibility to go into detail on a specific project event helped the<br />
participants in our study recognize it as an event worth including in their project<br />
timeline. Looking at project wikis, they provide reasonably good support for this e.g.<br />
through the page history overviews with links to each specific page [52]. Navigation<br />
in an email reader can be more cumbersome, but typically allows e.g. the use <strong>of</strong><br />
different windows for the mailboxes <strong>and</strong> the contents <strong>of</strong> individual messages. Instant<br />
messaging logs mainly allow navigation on a detailed level <strong>of</strong> project interaction only.<br />
With respect to overviews <strong>and</strong> navigation, limited features in the collaboration<br />
tools used to create the data may be amended by the use <strong>of</strong> other lightweight tools<br />
(e.g. search tools) to navigate the data. However, this comes at the loss <strong>of</strong> the<br />
contextualization <strong>of</strong> the data in the <strong>work</strong> (tool) environment from which it originated,<br />
a loss which may negatively affect participants’ sense-making <strong>of</strong> the data.<br />
Are the historical data subject to privacy issues? While the study reported in<br />
this paper considered data that were regarded as shared within the team, these<br />
characteristics do not apply to all data from project collaboration. IM logs, for<br />
instance, may be considered by their owners as too private to be shared [18]. <strong>The</strong>ir<br />
potential to support individual <strong>reflection</strong> on the project process may be considered.<br />
How will the use <strong>of</strong> historical data from the tool be integrated into a<br />
structured <strong>reflection</strong> activity with individual <strong>and</strong>/or collaborative sense making<br />
enabling participants to return to experience, reconstruct the stories [45] <strong>and</strong> draw<br />
lessons learned from the data? <strong>The</strong> organization <strong>of</strong> the activity can for instance be<br />
based on SD best practice for project retrospectives [8-9]. An issue which is not the<br />
main focus in this paper, but which is nevertheless essential <strong>and</strong> should be considered<br />
as part <strong>of</strong> the organization <strong>of</strong> retrospective <strong>reflection</strong>, is how to subsequently draw<br />
upon the lessons learned in the <strong>work</strong> practice, within <strong>and</strong>/or across projects <strong>and</strong> teams.<br />
For a project team considering the potential <strong>of</strong> one or more <strong>of</strong> their collaboration<br />
tools to support their <strong>reflection</strong> on, <strong>and</strong> <strong>learning</strong> from, their project process, it may be<br />
207
a good start to consider what aspects <strong>and</strong> challenges <strong>of</strong> their project <strong>work</strong> they would<br />
like to shed light on, <strong>and</strong> next consider what are the c<strong>and</strong>idate tools to contain<br />
relevant data.<br />
Finally, it should be stressed that when considering the potential <strong>of</strong> using a certain<br />
tool to aid retrospective <strong>reflection</strong> in a specific case, a general frame<strong>work</strong> is only a<br />
useful starting point. <strong>The</strong> functionality <strong>of</strong> the specific tool will have to be considered<br />
along with the usage <strong>of</strong> the tool in the specific project, which heavily impacts on what<br />
historical data is being stored through everyday use <strong>of</strong> the tool.<br />
8 Conclusion<br />
By investigating a case <strong>of</strong> retrospective <strong>reflection</strong> in a s<strong>of</strong>tware development<br />
project aided by historical data in an issue tracker (Trac), we have shown in detail<br />
how a collaboration tool in daily use in a project can be used to help participants learn<br />
from the project experience. We have demonstrated how certain types <strong>of</strong> data <strong>and</strong><br />
certain features for retrieving them proved particularly valuable to aid participants’<br />
recall <strong>and</strong> <strong>reflection</strong>, aiding the reconstruction <strong>of</strong> the project trajectory.<br />
Based on the findings we identified a set <strong>of</strong> questions that may be asked by a<br />
project team or a collaboration tool designer when considering the potential to use<br />
specific collaboration tools in retrospective <strong>reflection</strong>. <strong>The</strong> questions can serve as a<br />
starting point for the development <strong>of</strong> a more general frame<strong>work</strong> outlining issues that<br />
should be addressed. To gain insight on the potential <strong>of</strong> different types <strong>of</strong><br />
collaboration tools to aid retrospective <strong>reflection</strong>, more empirical studies are however<br />
needed. By unveiling project teams’ use <strong>of</strong> specific tools in retrospective <strong>reflection</strong><br />
we may identify ways <strong>of</strong> utilizing the tools that significantly differ from the way Trac<br />
was used in our study. This is an area <strong>of</strong> further research.<br />
Another issue to be addressed in further research is how a project team’s<br />
knowledge <strong>of</strong> the future use <strong>of</strong> a collaboration tool in retrospective <strong>reflection</strong> affects<br />
their daily use <strong>of</strong> the tool.<br />
Finally, we will examine how the retrospective use <strong>of</strong> Trac can be integrated into<br />
retrospective <strong>reflection</strong> <strong>work</strong>shops within a time schedule feasible for real SD<br />
projects. We will do this by trying out the approach in SD projects on a larger scale.<br />
References<br />
1. Carstensen, P.H. <strong>and</strong> K. Schmidt, <strong>Computer</strong> Supported Cooperative Work: New Challenges<br />
to Systems Design, in H<strong>and</strong>book <strong>of</strong> Human Factors/Ergonomics, K. Itoh, Editor. 2002<br />
(1999), Asakura Publishing: Tokyo.<br />
2. Grinter, R. Doing S<strong>of</strong>tware Development: Occasions for Automation <strong>and</strong> Formalisation. in<br />
ECSCW'97. 1997. Lancaster, United Kingdom: Springer.<br />
3. Herbsleb, J.D., et al. Distance, dependencies, <strong>and</strong> delay in a global collaboration. in<br />
CSCW'00. 2000. Philadelphia, PA: ACM<br />
4. Keil, M., J. Mann, <strong>and</strong> A. Rai, Why S<strong>of</strong>tware Projects Escalate: An Empirical Analysis <strong>and</strong><br />
Test <strong>of</strong> Four <strong>The</strong>oretical Models. MIS Quarterly, 2000. 24(4).<br />
208
5. Lyytinen, K. <strong>and</strong> D. Robey, Learning failure in information systems development.<br />
Information Systems Journal, 1999. 9: p. 17.<br />
6. Basili, V.R. <strong>and</strong> G. Caldiera, Improving S<strong>of</strong>tware Quality by Reusing Knowledge <strong>and</strong><br />
Experience. Sloan Management Review, 1995. Fall 1995.<br />
7. Dybå, T. <strong>and</strong> T. Dingsøyr, Empirical studies <strong>of</strong> agile s<strong>of</strong>tware development: A systematic<br />
review. Information <strong>and</strong> S<strong>of</strong>tware Technology, 2008. 2008(50): p. 833-859.<br />
8. Derby, E., D. Larsen, <strong>and</strong> K. Schwaber, Agile Retrospectives. 2006: Pragmatic Bookshelf.<br />
9. Kerth, N., Project Retrospectives: A H<strong>and</strong>book for Team Reviews 2001: Dorset House.<br />
10. Schindler, M. <strong>and</strong> M.J. Eppler, Harvesting project knowledge: a review <strong>of</strong> project <strong>learning</strong><br />
methods <strong>and</strong> success factors. International Journal <strong>of</strong> Project Management, 2003. 21: p. 10.<br />
11. Bjørnson, F.O., A.I. Wang, <strong>and</strong> E. Arisholm, Improving the effectiveness <strong>of</strong> root cause<br />
analysis in post mortem analysis: A controlled experiment. Information <strong>and</strong> S<strong>of</strong>tware<br />
Technology, 2009. 51: p. 150-161.<br />
12. Schön, D., <strong>The</strong> Reflective Practitioner. 1983: Basic Books, Inc.<br />
13. Kasi, V., et al., <strong>The</strong> post mortem paradox: a Delphi study <strong>of</strong> IT specialist perceptions.<br />
European Journal <strong>of</strong> Information Systems, 2008. 17: p. 62-78.<br />
14. Churchill, E.F. <strong>and</strong> S. Bly. It's all in the words: Supporting <strong>work</strong> activities with lightweight<br />
tools. in GROUP '99. 1999. Phoenix, Arizona, USA: ACM.<br />
15. Gutwin, C., R. Penner, <strong>and</strong> K. Schneider. Knowledge sharing in s<strong>of</strong>tware engineering:<br />
Group awareness in distributed s<strong>of</strong>tware development in CSCW'04. 2004. Chicago, Illinois,<br />
USA.: ACM Press.<br />
16. H<strong>and</strong>el, M. <strong>and</strong> J.D. Herbsleb. What is Chat Doing in the Workplace? in CSCW'02. 2002.<br />
New Orleans, Louisiana, USA: ACM.<br />
17. Krogstie, B.R. <strong>The</strong> wiki as an integrative tool in project <strong>work</strong>. in COOP. 2008. Carry-le-<br />
Rouet, Provence, France: Institut d’Etudes Politiques d’Aix-en-Provence.<br />
18. Krogstie, B.R. A model <strong>of</strong> retrospective <strong>reflection</strong> in project based <strong>learning</strong> utilizing<br />
historical data in collaborative tools. in EC-TEL 2009. 2009. Nice, France: Springer.<br />
19. Hutchins, E., Cognition in the Wild. 1995, Cambridge, Massachusetts: <strong>The</strong> MIT Press.<br />
20. Rogers, Y. <strong>and</strong> J. Ellis, Distributed Cognition: an alternative frame<strong>work</strong> for analyzing <strong>and</strong><br />
explaining collaborative <strong>work</strong>ing. Journal <strong>of</strong> Information Technology, 1994. 9: p. 119-128.<br />
21. Daradoumis, T. <strong>and</strong> M. Marques, Distributed Cognition in the Context <strong>of</strong> Virtual<br />
Collaborative Learning. Journal <strong>of</strong> Interactive Learning Research, 2002. 13(1/2): p. 135-<br />
148.<br />
22. Flor, N.V. <strong>and</strong> E.L. Hutchins. Analyzing Distributed Cognition in S<strong>of</strong>tware Teams: A Case<br />
Study <strong>of</strong> Team Programming during Perfective S<strong>of</strong>tware Maintenance. in Empirical Studies<br />
<strong>of</strong> Programmers: Fourth Workshop. 1991: Ablex Publishing.<br />
23. Sharp, H. <strong>and</strong> H. Robinson, Collaboration <strong>and</strong> co-ordination in mature eXtreme<br />
programming teams. International Journal <strong>of</strong> Human-<strong>Computer</strong> Studies, 2008. 66(7): p.<br />
506-518.<br />
24. Ackerman, M.S. <strong>and</strong> C. Halverson, Organizational Memory as Objects, Process, <strong>and</strong><br />
Trajectories: An Examination <strong>of</strong> Organizational Memory in Use. <strong>Computer</strong> Supported<br />
Cooperative Work, 2004. 13(2): p. 155-189.<br />
25. Salomon, G., No distribution without individuals' cognition, in Distributed Cognitions.<br />
Psychological <strong>and</strong> educational considerations, G. Salomon, Editor. 1993, Cambridge<br />
University Press.<br />
26. Strauss, A., Continual permutations <strong>of</strong> action. 1993, New York: Aldine de Gruyter.<br />
27. Grudin, J. <strong>and</strong> T. Lovejoy. Messaging <strong>and</strong> Formality: Will IM Follow in the Footsteps <strong>of</strong><br />
Email. in INTERACT 2003. 2003. Zurich: IOS Press.<br />
28. Isaacs, E., et al. <strong>The</strong> Character, Functions, <strong>and</strong> Styles <strong>of</strong> Instant Messaging in the<br />
Workplace. in CSCW'02. 2002. New Orleans, Louisiana, USA: ACM.<br />
29. Dourish, P. <strong>and</strong> V. Bellotti. Awareness <strong>and</strong> coordination in shared <strong>work</strong>spaces. in ACM<br />
CSCW. 1992. Toronto, Ontario, Canada ACM Press.<br />
209
30. Gutwin, C. <strong>and</strong> S. Greenberg, A Descriptive Frame<strong>work</strong> <strong>of</strong> Workspace Awareness for<br />
Real-Time Groupware. <strong>Computer</strong> Supported Cooperative Work 2002. 11(3-4): p. 411-446.<br />
31. Gutwin, C., R. Penner, <strong>and</strong> K. Schneider. Group Awareness in Distributed S<strong>of</strong>tware<br />
Development. in CSCW'04. 2004. Chicago, Illinois, USA: ACM.<br />
32. Bødker, S. <strong>and</strong> E. Christiansen, <strong>Computer</strong> Support for Social Awareness in Flexible Work.<br />
<strong>Computer</strong> Supported Cooperative Work, 2006. 15: p. 1-28.<br />
33. Schmidt, K. <strong>and</strong> C. Simone, Coordination Mechanisms: Towards a Conceptual Foundation<br />
<strong>of</strong> CSCW Systems Design. <strong>Computer</strong> Supported Cooperative Work, 1996. 5: p. 155-200.<br />
34. Johnson, J.N. <strong>and</strong> P.F. Dubois, Issue Tracking. Computing in Science <strong>and</strong> Engineering,<br />
2003(November/December ): p. 71-77.<br />
35. Prause, C.R., et al. Managing the Iterative Requirements Process in a Multi-National<br />
Project using an Issue Tracker. in IEEE International Conference on Global S<strong>of</strong>tware<br />
Engineering. 2008. Bangalore, India: IEEE.<br />
36. Poole, W.G. <strong>The</strong> S<strong>of</strong>ter Side <strong>of</strong> Custom S<strong>of</strong>tware Development: Working with the Other<br />
Players. in Conference on S<strong>of</strong>tware Engineering Education <strong>and</strong> Training. 2003. Madrid,<br />
Spain: IEEE.<br />
37. Kraut, R.E. <strong>and</strong> L.A. Streeter, Coordination in S<strong>of</strong>tware Development. Communications <strong>of</strong><br />
the ACM, 1995. 38(3).<br />
38. Whittaker, S., D. Frohlich, <strong>and</strong> D.-J. Owen. Informal Workplace Communication: What Is<br />
It Like <strong>and</strong> How Might We Support It? in Human Factors in Computing Systems. 1994.<br />
Boston, Massachusetts, USA: ACM Press.<br />
39. Dekel, U. <strong>and</strong> J.D. Herbsleb. Pushing Relevant Artifact Annotationsin Collaborative<br />
S<strong>of</strong>tware Development. in CSCW'08. 2008. San Diego, USA: ACM.<br />
40. Storey, M.-A., et al. Shared waypoints <strong>and</strong> social tagging to support collaboration in<br />
s<strong>of</strong>tware development. in CSCW'06. 2006. Banff: ACM.<br />
41. Fitzpatrick, G., P. Marshall, <strong>and</strong> A. Phillips. CVS integration with notification <strong>and</strong> chat:<br />
lightweight s<strong>of</strong>tware team collaboration. in CSCW'06. 2006. Banff, Alberta, Canada: ACM.<br />
42. Cubranic, D., et al. Learning from project history: a case study for s<strong>of</strong>tware development. in<br />
CSCW '04. 2004. Chicago, USA: ACM.<br />
43. Burge, J. <strong>and</strong> D.C. Brown. SEURAT: integrated rationale management. in ICSE'08. 2008.<br />
Leipzig, Germany: ACM.<br />
44. Burge, J. <strong>and</strong> J. Kiper. Capturing Collaborative Design Decisions <strong>and</strong> Rationale. in Design,<br />
Computing, <strong>and</strong> Cognition. 2008. Atlanta, Georgia, USA.<br />
45. Ar<strong>and</strong>a, J. <strong>and</strong> G. Venolia. <strong>The</strong> Secret Life Of Bugs: Going Past the Errors <strong>and</strong> Omssions in<br />
S<strong>of</strong>tware Repositories in International Conference on S<strong>of</strong>tware Engineering (ICSE'09).<br />
2009. Vancouver, Canada: IEEE.<br />
46. Trac home page, http://trac.edgewall.org/, Last accessed 10 September 2009.<br />
47. Krogstie, B.R. <strong>and</strong> M. Divitini. Shared timeline <strong>and</strong> individual experience: Supporting<br />
retrospective <strong>reflection</strong> in student s<strong>of</strong>tware engineering teams. in CSEE&T 2009. 2009.<br />
Hyderabad: IEEE <strong>Computer</strong> Society.<br />
48. Klein, H.K. <strong>and</strong> M.M. Myers, A Set <strong>of</strong> Principles for Conducting <strong>and</strong> Evaluating<br />
Interpretive Field Studies in Information Systems. MIS Quarterly, 1999. 23(1): p. 67-94.<br />
49. Wolf, C.G., J.R. Rhyne, <strong>and</strong> L.K. Briggs. Communication <strong>and</strong> Information Retrieval with a<br />
Pen-based Meeting Support Tool. in CSCW 1992.<br />
50. Star, S.L., <strong>The</strong> Structure <strong>of</strong> Ill-Structured Solutions: Boundary Objects <strong>and</strong> Heterogeneous<br />
Distributed Problem Solving, in Distributed Artificial Intelligence, M. Huhns <strong>and</strong> L.<br />
Gasser, Editors. 1990, Morgan Kaufmann Publishers. p. 37-54.<br />
51. Krogstie, B.R. Do's <strong>and</strong> dont's <strong>of</strong> instant messaging in students' project <strong>work</strong>. in NOKOBIT<br />
2009. 2009. Trondheim, Norway: Tapir.<br />
52. Krogstie, B.R. Using Project Wiki History to Reflect on the Project Process in 42nd Hawaii<br />
International Conference on System Sciences. 2009. Big Isl<strong>and</strong>, Hawaii: IEEE <strong>Computer</strong><br />
Society.<br />
210
Appendix B: Other research publications<br />
<strong>The</strong> following two research papers were published during the period <strong>of</strong> the PhD <strong>work</strong>.<br />
<strong>The</strong>y address student projects, but with a focus outside the scope <strong>of</strong> the thesis.<br />
Practice-Based Learning as Mobile Learning: <strong>The</strong> Role <strong>of</strong><br />
Boundary Objects<br />
Authors:<br />
Krogstie, Birgit <strong>and</strong> Divitini, Monica<br />
Published in:<br />
IADIS Mobile Learning 2007 , Lisbon, Portugal, 5-7 July.<br />
Abstract:<br />
Practice-based <strong>learning</strong> is an important element in study programs educating pr<strong>of</strong>essional<br />
practitioners. It is characterized by a focus on students’ transition from university to pr<strong>of</strong>essional<br />
<strong>work</strong> life. <strong>The</strong> students are exposed to real <strong>work</strong> challenges within the pedagogical scaffolding <strong>of</strong> the<br />
university program. An essential aspect <strong>of</strong> the <strong>learning</strong> process is students’ <strong>reflection</strong> on the<br />
relationship between school <strong>and</strong> real life, theory <strong>and</strong> practice.<br />
As we strive to provide the best possible pedagogical support for students’ <strong>learning</strong> in these<br />
educational settings, we look for the elements most likely to have a positive or negative impact on<br />
<strong>learning</strong> results. In this <strong>reflection</strong> paper, we will argue that a focus on interactional mobility is helpful<br />
in developing adequate support for practice-based <strong>learning</strong>, both in terms <strong>of</strong> course design <strong>and</strong>, more<br />
specifically, in terms <strong>of</strong> identifying collaboration technologies that may be helpful to the students. In<br />
particular, we focus on the role played by boundary objects <strong>and</strong> how scaffolding can be related to<br />
these objects.<br />
211
<strong>The</strong> <strong>work</strong>-<strong>reflection</strong>-<strong>learning</strong> <strong>cycle</strong> in SE student projects: Use <strong>of</strong> collaboration tools<br />
Learning from achievement: scaffolding student projects in<br />
s<strong>of</strong>tware engineering<br />
Authors:<br />
Bygstad, Bendik; Krogstie, Birgit <strong>and</strong> Grønli, Tor-Morten<br />
Published in:<br />
International Journal <strong>of</strong> Net<strong>work</strong>ed <strong>and</strong> Virtual Organizations, 6:2, 2009.<br />
Abstract:<br />
It has become almost a truism that students learn more from <strong>work</strong>ing on projects than from lectures.<br />
This is reflected in pedagogical approaches such as Problem-based Learning, Project based <strong>learning</strong><br />
(PBL) <strong>and</strong> Work-based Learning. A problem in PBL, underrated in the literature, is that while trivial<br />
tasks hold limited potential for <strong>learning</strong>, students may not succeed in solving nontrivial ones. We<br />
suggest that the solution lies in appropriate scaffolding: providing support for the learner to gradually<br />
master what is needed to complete a task. <strong>The</strong> empirical background for the study is a two-semester<br />
S<strong>of</strong>tware Engineering (SE) course at the Norwegian School <strong>of</strong> IT, with data collected over five years.<br />
We conclude that PBL in this setting may be successfully scaffolded by a formal, iterative <strong>and</strong><br />
incremental SE method. As our main contribution we point to six types <strong>of</strong> scaffolding addressing<br />
essential aspects <strong>of</strong> SE project <strong>work</strong>.<br />
212