--------
>
>Emanuel Knill writes:
>>
>>I have seen latex2html, and tried converting some of my
>>papers. Though I think it is great to
>>have build this tool, the results were not satisfactory for
>>what I tried. The most
>>obvious problem was symbol alignment.
>
Nikos Drakos writes:
>What is the symbol alignment problem?
>If you are referring to equation images not being aligned
>correctly within a piece of text that flows around them
>then this can be fixed with the new version of mosaic
>(XMosaic 2.0 prerelease 4) and the next release of latex2html
>(or the patch to v0.3.1 in
>http://cbl.leeds.ac.uk/nikos/tex2html/previous-versions/patch0.3.1-to-0.3.3.tx
>t)
>
The problem I was thinking of occurs with inline
greek and other math symbols. Their vertical alignment is
inconsistent. They hop all over the place (I assume because
the bottom of the image is alligned instead of the baseline
of the characters). I will try the patches.
Nikos Drakos continues:
>
>Apart from the macro capabilities of TeX, and the
>inclusion of style files there other problems
>resulting from the fact that in many cases single equations
>cannot be treated in isolation. For example automatic
>equation numbering, cross references using symbolic
>equation labels, automatic generation
>of theorem/subtheorem/axiom/lemma numbers cannot be
>addressed by embedding TeX in HTML unless parts of the
>TeX engine are replicated, or
>as Marc points out, unless TeX becomes more modular
>so that it can be used with other packages.
>For the time being these issues can only be addressed if
>the equations and the text around them are processed
>together and preferably off-line because of potential overheads.
>
Standard numbering in a math document designed primarily as
hypertext may not be what I would want... And you are right,
automatic generation of references across multiple documents
could be quite tricky, unless the original document is a
(La)TeX paper, in which case a conversion program like
latex2html makes sense.
Manny