The web version of each SEP document (a report or book) contains an HTML abstract, a Postscript version of the complete article or chapter, and a tar file with all the files that constitute the reproducible document directory. Furthermore, a web version of a report or book may contain the source of a program library or a data directory. We have three simple scripts to convert future SEP reports into our web format: paper2html, book2web, report2web. SEP reports earlier than report 70 are not reproducible and are represented by their ASCII source files.
We would prefer to display SEP documents in HTML rather than Postscript. Unfortunately, HTML does not yet support mathematical expressions. Since LATEX typesets mathematical expressions beautifully, SEP will continue to write its documents in LATEX. To transfer such documents to HTML routinely, SEP needs to maintain a latex2html enhancement for each of SEP's LATEX macros. Consequently, we suggest a drastic reduction in the number of available LATEX macros to the bare minimum. Furthermore, SEP may want a reliable mechanism to convert its LATEX documents to PDF format automatically.
As suggested above, the web and Java will probably change SEP's current reproducible document concept. A Java enabled browser and a document applet may replace the current combination of a dvi file, a dvi viewer, a figure's graphic user interface, and a GNU makefile. A web document's hyperlinks will be able to connect to a wide variety of local and remote resources. Java's dynamic loading of classes will possibly replace much of the makefile functionality.