Pablo Rodríguez wrote:
> Would your advice be the same couple of words ;-) if xmltex could be
> integrated with xelatex out of the box?
Yes, I think so. The problem with xmltex is
its limited ability to do context-sensitive rules
and access to the whole source.
Really, its a linear one-pass through the original.
I don't know about others here, but I have become
so used to the features of XSLT that I find it
hard to imagine life without them.
The TeX way of doing things (saving stuff in macros) reminds of
when we processed SGML with something like
David Megginson's Perl library - it's hard work.
And TeX's foibles with catcodes is a recipe
for going mad (eg when you process XML,
you set the character # to be inactive, then
you write to a file, read it back it, and lo! its active again)
> Do you think that the problem is xmltex or are TeX's boundaries the
> problem? I mean, do you think it is possible to develop a serious XML
> processor using TeX as typesetting engine?
Possibly. The XML-typesetting of ConTeXt is more
robust and sophisticated; but not that well
understood by most of us, sadly.
I'm open to being convinced, but I don't see the
work being done.
That's not to say TeX is not a good typesetting
engine behind the scenes.
> Which serious FOSS XML processors are available?
depends what you mean by "processors".
if you mean typesetting engines, I suppose
I would list
- Apache FOP XSL FO
I'd like to add Scribus, but I have no idea how
to get stuff into it.
> I would like to have something similar to the XML module in ConTeXt for
> XeLaTeX. I guessed this was xmltex.
no, definitely not. If you want that, you need to
persuade Hans Hagen to get an XeTeX-aware ConTeXt.
> XSL transformation to LaTeX source is fine, but it involves more than
> one step only for generating PDF output.
and I don't see that _per se_ as a problem. The difficulty
is that you end up writing parts of the system in
_radically_ different languages (eg XSL and LaTeX).
> Would you consider FOP or XEP professional XML processors? As far as I
> know they need XSL-FO. If XML and TeX are about separating content form
> layout, XSL-FO seems to be a too complex way to typeset texts and it
> would be not required by TeX.
> But if FOP or XEP are serious XML processors, the question I'm missing
> is what they have and xmltex does not.
FOP is not yet good enough (it does not implement all of XSL FO),
but all my experience with XEP (which is a lot) suggests that it
is fine. It implements FO properly, and FO is (more or less) capable
of describing serious typesetting.
Comparing FOP or XEP to xmltex is misleading. xmltex provides
both a language for writing rules about how to process XML,
and a typesetting engine. XEP and FOP simply implement a
page description language. All the intelligence is in the process
which converts your XML to XSL FO
> I don't know whether I understand the complexity of the issue. TEI
> (Lite) is by definition much more complex than its use by a single use
> or project. But xmltex might be useful to typeset particular projects.
yes, if your needs are simple, and your XML source
does not not need complex mangling, you could do
worse than xmltex. But remember that it's not xetex,
so it does not do Unicode properly, and its limited to
traditional TeX ways of doing fonts (ie you can't just
plug in an OpenType font)
> Even with XSL transformation into LaTeX source, adaptations are needed
> to typeset particular projects.
for what its worth (if its not obvious from previous postings),
I personally have reverted away from XSL FO back to transformations into
LaTeX, and processing using XeLaTeX. For two main reasons:
1. my own FO implementation (passivetex) was too limited,
FOP was/is incomplete, and XEP costs money
2. I am too familiar with handy LaTeX packages, and can't
face re-implementing them all in FO
as an example of the latter, I make PDF slides from TEI XML,
by transforming to LaTeX and using the "beamer" package.
No doubt I _could_ do that in FO, using XEP, but I just can't
find the time to work out how.
I suppose, though, my real reason is
3. I don't want/need to typeset as much as I did 20 years ago.
The web is the target these days....
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431