[This local archive copy mirrored from the canonical site: http://www.cis.ohio-state.edu/~gurari/temp/xml/ml.html; links may not have complete integrity, so use the canonical document at this URL if possible.]

XML and MathML from LaTeX using TeX4ht

The following are preliminary results of work done to configure TeX4ht for XML and MathML. The outcome demonstrates the potential of using TeX4ht with LaTeX (or other TeX-based systems) as a front end system for authoring XML documents.

A trivial HTML wrapper pointing to almost two hundred presentational MathML fragments results from latex to xml/mathml translation. The web page should render with techexplorer, subject to mistakes in the translation and missing features in the browser.
The math fragments are the <math> elements from this file, which is a 100% automatic translation of the LaTeX file x.tex. The XML file parses against latex.xmldtd (and other files referenced in there).
This is a non-MathML version compiled for arbitrary LaTeX-like math DTD latexmath.dtd.
The test file converted to normal HTML, with just the math converted to MathML.
The source LaTeX file x.tex, the configuration file xml.cfg, the virtual hypertext fonts cmex.htf, cmmi.htf, cmr.htf, cmsy.htf, ec.htf, and the DTD files mathml.dtd, latexent.dtd, latexent.dtd, latexentshort.dtd are offered for downloading.

The TeX4ht system, should be modified to include the more recent files tex4ht.c, tex4ht.perl, tex4ht.sty, and tex4hta.sty. The documentation in the home page, in the slides of a talk, and in a more recent web page, explain some aspects of the system, but some of the details are inaccurate.


The LaTeX dtd is not intended to be complete or finished; it simply implements enough to parse the test document. Similarly, we have not tried to implement all the TeX math commands properly, simply concentrating on those needed for initial tests.

This work is in preliminary stages. We would like to hear from anyone interested in testing it further.

Eitan Gurari
Sebastian Rahtz
August 7, 1998