[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: mathematics and paraplegia

Karl Winterling <kwinterling@gmail.com> wrote:
> Yeah, some alternatives (like MathML and Lout) are emerging, but you
> should stick to LaTeX for anything serious until those are well
> supported.

Actually, the author of Lout has moved on to a different (still free software,
as far as I know) typesetting project. I don't know how far it has progressed,
but it's worth watching.

XML has some advantages, mostly inherited from SGML. Unfortunately, MathML is
horribly verbose syntactically, so the only reasonable way to create it is to
write the mathematics in TeX or a similar format and then convert it to
MathML. ITeX2MML is a simple translator that does this from a TeX-like format.

As to typesetting, it is possible to start with XML, or a format such as
RestructedText or one of the simple wiki-based markup languages from which XML
can be derived. The typesetting process then usually involves the conversion
of the XML to TeX macros, or alternatively, direct conversion to PDF via Fop,
which is still under development and intended to implement the full XSL
specification. If the typesetting is to be accomplished by TeX anyway, then I
don't see much advantage in starting with XML rather than writing LaTeX in the
first place, especially when one considers that LaTeX can be converted to
XHTML+MathML with tex4ht.

Thus I don't think there will be a viable TeX substitute until a new
typesetting system becomes available or Fop becomes compelling. If you only
need to deliver some form of HTML to the Web, then writing HTML directly, or
RestructredText or a similar format, would be best in my opinion.

Most of my writing is not mathematical in nature, but it is primarily intended
to be printed or viewed in typeset form as PDF files. This is why I prepare it
in LaTeX.

On the accessibility side, returning to the topic of this thread, formats such
as RestructuredText could be easier to edit than XML, particularly for users
of alternative input devices, although again, the Emacs nxml mode offers
validation and text completion for element and attribute names, so as long as
you don't need to write MathML XML might not pose any serious difficulties.

Reply to: