28.06.2013 Views

Papers in PDF format

Papers in PDF format

Papers in PDF format

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

would not have the rights to change a paper they are referee<strong>in</strong>g, but they would have the rights to make annotations. If<br />

desired the author could get the right to also annotate the referees' annotations to clarify misunderstand<strong>in</strong>gs - naturally<br />

both referees and author rema<strong>in</strong><strong>in</strong>g anonymous.<br />

For the whole process from adm<strong>in</strong>istration to f<strong>in</strong>al document preparation a set of tools and fill-out forms has been<br />

developed that can be adopted accord<strong>in</strong>g to the publishers' needs. The follow<strong>in</strong>g paragraphs deal with some of the more<br />

sophisticated tools that have been developed to provide a cost effective way for publishers to add value to electronic<br />

papers.<br />

As mentioned above it is desirable to keep older versions of documents for archival and ma<strong>in</strong>tenance purposes. For this<br />

reason a special version<strong>in</strong>g tool was developed that uses Hyper-G's cluster mechanism: different versions of a document<br />

are clustered together and the reader can switch back and forth between different versions on the fly. Whenever a new<br />

version of a document is <strong>in</strong>serted a special l<strong>in</strong>k migration tool parses all outgo<strong>in</strong>g hyperl<strong>in</strong>ks <strong>in</strong> the older document and<br />

<strong>in</strong>serts them at the appropriate location <strong>in</strong> the new document. Only outgo<strong>in</strong>g l<strong>in</strong>ks are considered dur<strong>in</strong>g migration<br />

because <strong>in</strong>com<strong>in</strong>g l<strong>in</strong>ks should not be touched - they are normally comments for exactly the version of the paper they<br />

are po<strong>in</strong>t<strong>in</strong>g to.<br />

Consider<strong>in</strong>g a paper one will f<strong>in</strong>d a lot of so-called vocative l<strong>in</strong>ks. A vocative l<strong>in</strong>k is a textual po<strong>in</strong>ter to a location such<br />

as ``see also page nn''. In scientific papers one will normally f<strong>in</strong>d a references section with po<strong>in</strong>ters to other publications<br />

- aga<strong>in</strong> a typical example for vocative l<strong>in</strong>ks.<br />

These considerations about vocative l<strong>in</strong>ks have lead to the development of an automatic vocative l<strong>in</strong>k creation tool. The<br />

tool is based on a so called ``Vocative L<strong>in</strong>k Creation Language'' (VLCL) that was designed to support the description of<br />

contexts <strong>in</strong> documents and to f<strong>in</strong>d potential hyperl<strong>in</strong>ks <strong>in</strong> it. To not loose myself <strong>in</strong> too many details, here is a small<br />

example:<br />

Consider a piece of text with the phrase ...details can be found <strong>in</strong> [Moser 95].... The tool will identify this phrase to be a<br />

vocative hyperl<strong>in</strong>k lead<strong>in</strong>g to the references section and <strong>in</strong>sert a l<strong>in</strong>k to the references here. Pars<strong>in</strong>g the references it<br />

f<strong>in</strong>ds the entry [Moser 95] Moser J., ``The Art of PostScript programm<strong>in</strong>g'', available at http://www.iicm.edu/app, Dec.<br />

95. The URL will be identified by the tool and a l<strong>in</strong>k to the accord<strong>in</strong>g document will be <strong>in</strong>serted. S<strong>in</strong>ce VLCL is a<br />

programm<strong>in</strong>g language the behaviour can even be controlled to the extent that the l<strong>in</strong>k to the references is not created<br />

but leads directly to the location <strong>in</strong>stead.<br />

If this sounds complicated - VLCL was designed to describe such rules <strong>in</strong> a very compact and high level manner.<br />

Normally VLCL programs hav<strong>in</strong>g the functionality described above are not longer than e.g. 30 -- 40 l<strong>in</strong>es. And they<br />

normally have to be written only once because journals have their well def<strong>in</strong>ed citation rules that are not chang<strong>in</strong>g too<br />

much over time. In addition citation rules are not vary<strong>in</strong>g too much between journals so that an exist<strong>in</strong>g VLCL program<br />

can be slightly modified and will then suit the needs of another journal.<br />

A different tool for automatic hyperl<strong>in</strong>k creation deals with glossaries. A glossary <strong>in</strong> Hyper-G is def<strong>in</strong>ed as an arbitrary<br />

collection of explanatory documents that are classified by their titles and keywords. The glossary l<strong>in</strong>k creation tool<br />

accepts arbitrary many glossary collections and automatically <strong>in</strong>terl<strong>in</strong>ks items <strong>in</strong> papers with glossary entries. The l<strong>in</strong>ks<br />

created get the special type glossary so that the reader can turn them on and off seperately when needed. The glossary<br />

l<strong>in</strong>k creation tool works at the moment for HTML documents, <strong>PDF</strong> and PostScript support are under development.<br />

Feedback about the readers' behaviour is necessary to improve the quality of a server. Standard statistics tools today are<br />

able to count the number of accesses to a document and give <strong>in</strong><strong>format</strong>ion about the location of the reader. Due to<br />

Hyper-G's user session concept a lot more <strong>in</strong><strong>format</strong>ion can be extracted from the logfiles. For this reason a specialized<br />

statistics tool was developed that is also able to trace the readers' way through the server from the beg<strong>in</strong>n<strong>in</strong>g to the end<br />

of a session.<br />

Us<strong>in</strong>g the session-oriented statistics the <strong>in</strong><strong>format</strong>ion provider can see the ``typical'' way of a reader through the server<br />

and f<strong>in</strong>d out about specific problems that arise. As an example the server operator could see that a majority of users is<br />

quitt<strong>in</strong>g a session when download<strong>in</strong>g a certa<strong>in</strong> document. In this case the <strong>in</strong><strong>format</strong>ion provider could have a look at this<br />

document and e.g. f<strong>in</strong>d out that there is a huge <strong>in</strong>l<strong>in</strong>e image on this page that takes to much time to be downloaded.<br />

Another situation could be that l<strong>in</strong>k references are misunderstood and the readers get a page they don't want.<br />

It can also be analysed if the structure on the server is easy to handle for the reader. Items that are often searched <strong>in</strong>stead<br />

of accessed directly are very likely not reachable easy enough. Parts of the structure that are never accessed also alert the

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

Saved successfully!

Ooh no, something went wrong!