13.07.2015 Views

Reviews - MetaLab

Reviews - MetaLab

Reviews - MetaLab

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Reviews</strong>Presented by:Lynn Tan Shun Chee (SW087612)Yogaraj A/l Rajendran (SW088209)Givita A/p Rama Roa (SW087602)Sathish Kumar A/l Rajoo (SW087642)


Peer Review


Participants of Peer <strong>Reviews</strong>• 3 – 5 members• Review Leader• Author• Specialized professionals• Designer• Coder or Implementer• TesterInspections• Standards Enforcer• Maintenance Expert• User RepresentativeWalkthroughs


Preparation for a peer review session• Review Team Leader & Author• Which sections to be reviewed??• Who should be the team members??• How long will the review take??• Documents• Inspections• Read and comments before begins• Overview Meeting• Walkthroughs• Read and comment in general overview


Peer Review Session• Walkthroughs• Read the documents and adds if needed.• Comment and explanation if necessary.• Inspections• Short presentation or overview of projects• Design sections to be reviewed.• Scribe should classified according to severity.• Participant• Comment and interrupt.PresenterAuthorsBoth


Session Documentation• Walkthroughs• Walkthrough session findings report• Inspections• Inspection session findings report• Inspection session summary report• Handed immediately after the session end.


Post-peer Review Activities• Prompt, effective correction and reworking of all errors• Follow up activities• Transmission of the inspections report to the internal Corrective ActionBoard (CAB) for analysis.• Prevention actions that reduce future defects• Improve productivity.


Peer Review CoverageSections recommended for inclusionComplicated logicCritical sectionsSections dealing with newenvironmentSections designed by new orinexperienced new membersSections recommended for omissionStraightforward sectionsSections that already reviewed severaltimes by similar past projectsSections that will not affectfunctionalityReused or repeated design and code


Efficiency of peer reviews• Peer review detection efficiency• Average hours worked/defect detected• Peer review defect detection density• Average number of defects detected/page of the design document• Internal peer review effectiveness• Percentage of defects detected by peer review as a percentage of total defectsdetected by the developer.


Efficiency of peer reviews• Dobbins(1998) quotes Madachy’s findings from an analysis of the designand code inspections conducted on the Litton project.• Madachy’s findings regarding the first two metrics cited above arepresented in table.• Dobbins(1998) also cites Don O’Neil’s 1992 National Software QualityExperiment.• It is conducted in 27 inspection laboratories operating in the US.• A total of 90925 source code lines were code-inspected


Efficiency of peer reviewsType ofdocumentNo. ofinspectionTotalnumber ofdefects andmajordefectsNo. ofpagesInspectionresourcesinvested(workhours)Defectdetectiondensity(defects/page)Inspectiondetectionefficiency(workhours/majordefect)RequirementdescriptionRequirementanalysisHigh-leveldesignTestprocedures21 1243(89 major)32 2165(117major)41 2398(198 major)18 1495(121 major)Code 150 7165(772 major)552 328 2.25 3.691065 769 2.03 6.571652 1097 1.45 5.571621 457 0.92 3.785047 4612 1.42 5.97DesignInspectionCodeInspection


Efficiency of peer reviews• Total number of defects detected 1849• Number of major defects detected 242• Total preparation time(min) 22828• Accordingly:a) Average preparation time per detected defect12.3min(0.2hr)b) Average preparation time per detected major defect94.3min(1.57hr)


Efficiency of peer reviewsTotal defect detectiondensity (defects perKLOC*)Major defect detectiondensity (defects perKLOC*)National SoftwareQualityLitton project20.3 25.92.66 2.80Comparison of defect densities detected between National Software Quality and Litton project


Efficiency of peer reviews• The internal effectiveness of inspections is discussed byCusumano(1991).• He also reports the results of a study on the effectiveness of designreview ,code inspection and testing at Fujitsu(Japan)(1977-1982).• A comparison by year of inspection , presented in table.


Efficiency of peer reviewsYear Defect detection method Defects perTest% Design Code 1000 lines ofreview% inspection% maintainedcode1977 85 - 15 0.191978 80 5 15 0.131979 70 10 20 0.061980 60 15 25 0.051981 40 30 30 0.041982 30 40 30 0.02Code Effectiveness at Fujitsu according to Cusumano


Inspection vsWalkthrough


Process of inspection1) Organizational preparation2) Overview meeting3) Thorough review of document4) Inspection session5) Corrections and reworking6) Follow up of correction and reworking


Process of Walkthrough1) Organizational preparation2) Brief overview reading3) Walkthrough session(s)


Comparison of the teamreview methods


Properties Formal design reviews Inspections WalkthroughsMain direct objectives-Detect errors-Identify new risks-Approve the designdocument-Detect errors- Identify deviationsfrom standards-Detect errorsMain indirect objectives -Knowledge exchange -Knowledge exchange- Support correctiveactionsReview leaderParticipantsProject leaderparticipationChief software engineeror senior staff memberTop-level staff andcustomerrepresentativesTrained moderator(peer)Peers-Knowledge exchangeCoordinator (peer , theproject leader onoccasion)PeersYes Yes Yes: usually as thereview’s initiatorSpecializedprofessionals in theteam- -Designer-Coder or implementer-Tester-Standards enforcer-Maintenance expert-User representative


Process of reviewFormal designreviewsInspectionsWalkthroughsOverviewmeetingNo Yes YesParticipantspreparationsYes-thorough Yes-thorough Yes-briefReview session Yes Yes YesFollow-up ofcorrectionsYes Yes No


InfrastructureFormal designreviewsInspectionsWalkthroughsFormal training ofparticipantsNo Yes NoUse of checklists No Yes NoError-related datacollectionNot formallyrequiredFormally required Not formallyrequiredReviewdocumentationFormal designreview report-Inspectionsession findingsreport-Inspectionsession summaryreportWalkthroughsession findingsreport


Expert opinions


Expert opinion• Prepared by outside experts• Support quality evaluation


Duties of Expert Opinion• Preparing expert judgment about documents.• Participating as a member of internal design review,Inspection or walkthrough team.


Benefits of Expert opinion• Insufficient in-house professional capabilities in specialized area• Lack of in-house professional due to workload pressures.• In small organizations, where the number of suitable candidates forreview team is insufficient.


Activity Time


Is time to role play• Presenter• Choose a part of design view to be discuss.• Inspection• Drop down error according to severity.• Walkthrough (standards enforcer, maintenance expert)• Drop down error according to your role.• Walkthrough (user)• Drop down error according to your role.• Expert Opinion• Give the expert judgment


Any Question??


Quiz Time


1. Name the 4 direct objectives of review.2. Name the 2 documents to be produced after an inspection session.3. What is the task for review leader?4. Why a review team is needed in software development ?5. What is the task for development team ?6. Name the two types of the peer review.7. For walkthrough what are the recommended professionals?8. State 4 Design Review Session Agenda9. In Pressman’s 13 “Golden Guidelines”. There is a collection of 3subgroup in the guidelines. State the subgroup and give one of the“Golden Guidelines” for each subgroup.


10. Examples of sections that should be and should not be included inpeer review.11. In overview meeting there are 6 types of agenda occur. List 3 agenda?12. What is the role of walkthrough coordinator?13. List the participants for formal design review, inspections andwalkthrough.14. List down the main indirect objectives of formal design reviews,inspection and walkthrough15. What is the duty of expert opinion


Thank you andGood Luck

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!