Re: [xml-dev] SGML complexity

[<juanrgonzaleza@canonicalscience.com>]

Philippe Poulard said:

>
> I'm not sure that client side Javascript is also the best language to
> parse LaTeX and create a MathML document.

It is not. That is the reason for E4X EcmaScript standard, but currently
is only supported by Firefox 1.5, therefore i did not explore this way.

>> And sure also i could do full match pattern transforming predefined
>> equations and store several billions of predefined patterns in an
>> external file.
>
> What kind of programmer would consider such an ugly idea ?

A desperate XSLT hacker? ;-)

>
> Maybe your own tool ? You are telling again and again that XSLT is not
> suitable for such transformation, I do really agree on this point.
>
> Others here :
> http://www.google.fr/search?hl=fr&q=latex2xml
>

Thanks by the link. It is true that you can write a XSLT transforming
textual content as \alpha to XML stuff, for instance to <mi>&alpha;</mi>,
and you can write 20 XSLT matching rules. The XSLT is very verbose but
works. However, as correctly pointed Carlisle already matching braces is a
interesting issue in XSLT.

I have taken a look to documentation of the latex2xml project and whereas
you can see TeX commands as \"o, \scriptsize, \Omega, and \$, ^, \}, or \(
between others in the LaTeXCommands.xml file, you do not see simple { or }
or \frac{}{}. Probably, the preparser transforms matching braces and other
difficult stuff and after the remaining text is done via XSLT. I do not
know.

> Cordialement,
>
>                ///
>               (. .)
>   --------ooO--(_)--Ooo--------
> |      Philippe Poulard       |
>   -----------------------------
>   http://reflex.gforge.inria.fr/
>         Have the RefleX !


Juan R.

Center for CANONICAL |SCIENCE)