W3C home > Mailing lists > Public > www-math@w3.org > January 2006

Re: mathml applications

From: Bruce Miller <bruce.miller@nist.gov>
Date: Mon, 09 Jan 2006 10:54:41 -0500
Message-ID: <43C28741.8000906@nist.gov>
To: Neil Soiffer <neils@dessci.com>
CC: Public MathML mailing list <www-math@w3.org>

Neil Soiffer wrote:
> This week, I will be giving an update to a talk I gave last year about 
> MathML applications.  If anyone has some new applications not listed on 
> http://www.w3.org/Math/Software/ or if you have an update to info listed 
> there, please send that info so I can incorporate it into the talk.

There's my LaTeXML, which I've been meaning to submit to the software
list, once I get a breath.
See  http://dlmf.nist.gov/LaTeXML/
for description, documentation and a (somewhat out-of-date) example from DLMF.

A brief description (and apologies, in advance, for beating my own drum):

LaTeXML attempts to mimic TeX's behaviour as fully as possible, but interprets
TeX via a Perl program, rather than using TeX's engine itself.  Thus extension
and customization can be done using perl modules, as well as TeX code.

It processes whole documents (although it can be induced to process fragments,
such as formula), converting the document to a LaTeXML DocType that corresponds
to LaTeX's constructs.  This avoids the information lost by direct conversion
to html. This output can then be converted to html (w/images for math),
xhtml (w/MathML) or other formats using XSL stylesheets (included for html/xhtml).

[In principle, but not practically, it could also be made to directly generate
 other XML formats: Ie. the processing engine is (almost) separated from the
 document generation. However, since the vocabularies are so large it's nontrivial
 to replace the generation definitions.]

As for MathML: the program attempts to parse the formula into, at least,
a parse tree.  This is necessary to generate "Good" presentation MathML.
It is close to enough to generate content MathML, as well, although
some major sticking points and ambiguities, such as the meaning of superscripts
and the handling of unknown symbols, are the subject of further development.
To increase the quality of the parsing, an author can either use special
markup to disambiguate the notations, or use document-specific external declarations.


> The audience will likely have a strong TeX background.  If you have had 
> any experience (pro or con) with the various TeX-to-XHTML+MathML 
> converters in the last year, I'd like to hear about that so I can pass 
> it along.

I've been using it pretty heavily in the DLMF (Digital Library of Mathematical
Functions) project.  I've processed 26 chapters (must be equivalent to
6-700 pages) so far; we're preparing a draft of the site for evaluation
(alas, the editors are keeping it close to the chest, however).
So, I'd have to say my experience is positive :>

Michael Kohlhase has also been using LaTeXML in developing his STeX system.

> Thanks,
>  
> Neil Soiffer
> Senior Scientist
> Design Science, Inc.
> www.dessci.com <http://www.dessci.com>
> ~ Makers of Equation Editor, MathType, MathPlayer and MathFlow ~


-- 
bruce.miller@nist.gov
http://math.nist.gov/~BMiller/
Received on Monday, 9 January 2006 15:54:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:58 GMT