For the math part of (La)TeX, one idea that has been used successfully
is to use XML delimiters for math mode (such as < m > or < me > instead
of $ and $$ or \(\) and \[\] ), and leave the actual math in (La)TeX
markup. That keeps it human-readable, and no information has been lost.
Later you can convert just the math, in the unlikely event that you want
that content in MathML.
That is the approach used by the PreTeXt authoring markup language:
ps. You do have to convert the & and < in math mode, to \amp and \le .
On Tue, 7 Aug 2018, Rosanna Cantavella wrote:
> Thank you so much, Gerrit and Peter. What a luxury is this list, as I'm
> being responded by the best!
> OK, I can see that, although there are no miracles yet, the process has
> more possibilities than it had some years ago.
> We'll study your excellent proposals (thanks again for them!) and I'm
> sure we'll be able to make a start from here on these texts.
> Best wishes,
>> On 06/08/18 12:46, Rosanna Cantavella wrote:
>>> Dear list,
>>> Would anybody kindly tell me if advances in automatic conversion of
>>> files from TeX to TEI-XML have been made?
>> Not that I am aware of, and certainly not for plain TeX. For
>> LaTeX-to-XML, see
>> http://latex.silmaril.ie/formattinginformation/latexto.html and the note
>> about Pandoc.
>> Iff the LaTeX is well-formed (no custom macros, no redefining of
>> primitives, etc) and uses the standard default commands as described in
>> the LaTeX Book, then Pandoc will do an excellent transformation to a
>> variety of XNL formats including TEI Simple.
>> However...(you saw that coming :-) most LaTeX is anything but
>> well-formed in that sense. On the VERY rare occasions when I have to do
>> this (and I have one pending at the moment), I use a combination of
>> Emacs macros, global replaces, Linux shell scripts, and maybe Pandoc.
>>> We would be interested in converting a synoptic edition of the
>>> complete works of the Catalan poet Ausiàs March (many thousands of
>>> lines on many witnessess) using this process.
>> If it is regularly-formed (does things the same way every time) then it
>> is certainly possible.
>> One route you might want to investigate, if the volume is large, and if
>> there is some money available, is using one of the many companies in the
>> Pacific Rim who do this kind of conversion.
> Prof. Rosanna Cantavella
> Universitat de València
> Life Member, Clare Hall University of Cambridge
> Editor, Magnificat Cultura i Literatura Medievals