4 Conclusions

In this paper we have described a system that handles the automatic hypertext conversion and representation of printed documents. We have illustrated some of the implemented functionality that is used in order to provide convenient access to the WWW version of the hypertext base. Many parts of this functionality have been implemented using or extending commercial (e. g. SCANWORX from Xerox Imaging Systems) or publicly available software (e. g. PBMTOOLS by Jef Poskanzer for image conversion). This approach has been chosen in order to prove feasibility. Therefore, still many parts of the WWW server scripts may be optimized, especially with regard to the speed of creating dynamic representations of pages i. e. with regard to the dynamic creation of temporary image files used for highlighting or zooming.

However, some parts of the WWW server scripts cannot be adjusted at all (or requiring a lot of effort) to the original HYPERFACS system's functionality. This is mainly due to different models used in WWW and HYPERFACS. E. g. as already mentioned, manual link generation, as supported in HYPERFACS, is hard to remodel in WWW. Another example of functionality that is hard to implement on the basis of WWW at the moment, is the creation and dynamic selection of document sets (cf. [4]). Therefore, the full text retrieval component of our server takes exactly one database or data file to search, instead of searching within a dynamic selection of databases that may be indicated by the user. Such a search mechanism would be especially helpful within a complete library of documents.

Nevertheless, by means of WWW and the MOSAIC browser, it was possible to create a sophisticated access system to image representations of documents within a restricted amount of time. Such a system could be advantageously used e. g. in a library or in a publishing environment [7]. In both cases, clients could have access to centralized and up-to-date documents from their desktops making use of advanced search facilities. However, clients without network access could also make profit from these mechanisms: branches of libraries could install public access stations and publishers could distribute additional electronic versions of their publications. The still increasing dissemination of MOSAIC and the WWW will ensure the needed dissemination of knowledge as well as browsing tools [2].