From one source to many documents

We wrote three relatively short Perl scripts and a few make rules to enable the author to translate his document file, paper.tex, to his favorite output format, such as Postscript, HTML, DVI, etc. The three Perl scripts - Actify, Latify, and latex2dvi - are extensible and easily maintained. These scripts replace SEP's former C-shell script, texpr.

Actify

extracts relevant information, such as RESDIR (the result directory) and ER, CR, NR (reproducibility status), from the makefile and includes it in the document file. Until now, authors had to state the RESDIR in both the document file and the makefile. The reproducibility status was clumsily communicated via the creation of dummy files.

Latify

turns the document source file into a complete L^ATEX file. Latify adds a document style statement, a begin and end document statement, etc. The script's output is a standard L^ATEX file that uses SEP macros.

latex2dvi

translates standard L^ATEX to a device independent format (DVI). Basically, it runs latex, bibtex, makeindex as many times as necessary.

This set of Perl scripts currently produces a family of output formats of the original document: The most common formats are DVI versions, that can be displayed with a DVI reader (xdvi) or converted to Postscript (dvips). For the time being, we additionally support an enhanced DVI format based on the xtex DVI reader. Xtex permits us to display DVI files that have buttons in the figure captions. When pressed, these buttons bring up a graphics user interface that enables the reader to interact with the figure. We hope to replace xtex by an alternative interactive renderer, since xtex is non-standard and hard to maintain.

Additionally, we use the freely available latex2html script, which translates standard L^ATEX to HTML. We enhanced the latex2html Perl script to process certain non-trivial SEP macros into their HTML implementations. In particular, SEP's figure macro activeplot is translated by a customized Perl subroutine (latex2html explicitly endorses such enhancements.) In the next section, we discuss our attempt to create interactive HTML pages for SEP's reproducible documents.

Additionally, we are currently testing an alternative web solution based on the Java DVI renderer, IDVI (<http://www.geom.umn.edu/java/idvi/>). IDVI allows a reader to interact with a document through embedded Java applets.