W3C home > Mailing lists > Public > www-math@w3.org > October 1998

Re: Literate programming in MathML. Was: STATISTICS

From: David Carlisle <davidc@nag.co.uk>
Date: Fri, 16 Oct 1998 10:42:36 +0100 (BST)
Message-Id: <199810160942.KAA14540@nag.co.uk>
To: rafael.amen@ivf.se
CC: www-math@w3.org

>  I've been using Gurari's ProTeX
> (http://www.cis.ohio-state.edu/~gurari/systems.html)
> with a special extension, so that math formulae are metaencoded once
> and then output 
> * as typeset documentation (TeX quality) and 
> * as program code (Fortran) 

You may be interested in Eitan Gurari's tex->mathml system

> with two XSL sheets: the default one 
> for documentation and a special one for code output
> My question is: Is it possible to create such a double XSL sheet system

Unfortunately not yet with XSL, as the current XSL version 1 draft has
no support for mathematics. (It explicitly states that Math will be
looked at in later drafts.) To see what one might do with MathML and
stylesheets, I produced some DSSSL scripts for MathML, available

However the _principle_ of having multiple stylesheets, once XSL has the
needed features is I think sound. (One could probably do the special one
for code output already, as it doesn't require the math formatting
objects to be defined in XSL).

>  If it is, which markup tag set should the 
> code XSL sheet be based on? Why?

On the grounds that the more information that is preserved the better,
probably you should use MathML Content markup if you have that semantic
information available. Note that Content Markup does provide a default
rendering. If that is unsuitable for any particular form you can always
override the rendering for particular expressions by attatching
Presentation MathML expression via a semantics element or definition.
(This means of course that your stylesheet needs to be able to format
both Presentation and Content MathML, however my experience with the
DSSL scripts above is that this is not too hard (although I haven't yet
implemented it all:-).

Received on Friday, 16 October 1998 05:45:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:28 UTC