Re: comments re draft version 2.0

> There is no such thing as TeX syntax for Content
why ever not?

A tex package could fairly easily be written that had exactly the
constructs of content mathml, but in more traditional tex syntax.
\apply{\sin \ci{x}}
for instace, being a syntax I just made up, but which is clearly
transformable to and from
<apply><sin/> <ci>x</ci></apply>
with no loss of information.

> , except in so far as TeX-the-programming-language could
> masochistically be used to express it  

It's fairly easy to get tex-the-program to read arbitrary xml, for
example my xmltex system which is a fairly complete xml 1.0 + XML
namespace parser is written in tex and typesets a growing subset of
mathml, XSL FO, docbook and TEI flavours of XML.

> TeX-the-markup-language is a Presentation representation.
> TeX to Presentation translation is a solved problem (e.g. TtM).

There are several systems do do tex to mathml, I've still had more
success with tex4ht than anything else and given arbitrary tex macro
markup a system like tex4ht which uses tex to do the macro expansion
will always have the edge over a system trying to parse tex with
an external parser, I fear. I don't think any of the systems are
sufficiently reliable that the problem could be classed as completely
solved so far, unless things have changed a lot recently? However
I agree that this side of things is almost there and tools will
soon be available to do reliable batch conversions of  legacy documents.


> TeX to Content translation, I submit, is an exercise in futility, for the
> simple reason that there is no unique translation from Presentation to

> Content. The semantics aren't defined.  

If the TeX macro package is specifically designed to encode the
semantics of Content MathML, then the exercise is not at all futile.
This is, as I commented in my message that you quoted, a distinct
activity from converting legacy documents.


David

Received on Tuesday, 11 April 2000 18:05:23 UTC