39.5 Producing the Hypertext for the Web

Compared to the DAGS'92 proceedings, the hypertext part of DAGS'95 required much less postprocessing effort. This was mostly due to the fact that authors got an HTML template well before actually turning in the final version of their paper. All the authors that submitted their paper in electronic format did respect our editing guidelines. The only activity on our side was to actually link the HTML files into the DAGS'95 proceedings directory structure. While the DAGS'92 proceedings offer much more hypertext functionality, this functionality came at the price of manual postprocessing, resulting in much higher production cost and a longer production cycle. Shifting the burden of hypertext editing to the author relieves the editors of hypermedia proceedings of their biggest task. The price that we had to pay for DAGS'95 is the incompleteness of the proceedings, as we had no resources to manually convert to HTML the papers that we only got on paper; using OCR technology as we had done it for the DAGS'92 proceedings would have overstretched our project budget.

The structure of a DAGS'95 paper on the Web is less than ideal in that each paper is implemented as one, large HTML file with a table of contents in the beginning. This structure makes it harder to add links within a paper and to offer hierarchical navigation as available in the Gloor/Dynes hypertext engine. Additionally, downloading one large text file may take some time on a slow network connection. Nevertheless, our experiment shows that publishing proceedings on the web offers a valuable extension of printed proceedings at much lower cost than CD-ROM publishing.