Chris Rowley writes:
 > > internally the PDF model is simplistic - but then who *has*
 > > implemented anything better in mainstream software?
 >
 > As a model of what is it simplistic?
as a model of hypertex requirements. it does not permit support
targets, for instance

 > Some of it is the opposite of
 > simplistic, being over-speciliased and baroque.
thats just the syntax, i assume?

 > And what are the `internals of a model' in this context?
the features it supports, as opposed to the user syntax

 > But perhaps he just mean its model of links and other hyper-stuff?
yes

 > And even more important to implement the infrastructure to make
 > something like pdfLaTeX (eg context;-) able to easily, portably and
 > transparently convert logical browser formats into well-formatted and
 > well-typeset documents (for both screen and paper) described in some
 > sufficientlly rich, device-independent language.
oh, you've read the XSL spec then?

 > > by which we see why LaTeX is unpopular in production workflows. that
 > > translates to "10% failure"
 >
 > 10% failure would be heaven in our production typesetting environment
if they get more than 10% failure, why in heaven do they persist? it
sounds like total madness

 > (and we do not see any efficiency gain from sending the stuff
 > across the world to be keyboarded ... but this probably
 > short-sighted).
probably it is. a typical data entry firm will get you *very* high
quality useable eg SGML files. predictable cost, predictable
processing. no more catcodes.

 > But we still have enough people around who recall the problems we used
 > to have with 30% failure in a galley/paste-up hard-copy external
 > typeseting system: even 40% failure with electronic typesetting/editing
 > is, for them, absolute zen already!
poor souls
 > > a new language, using XML syntax, to say whatever you want
 > >   <foo n="3">x<bar>y</bar></foo>
 > > (forget the verbose syntax for now), and then provide the XSL
 > > transformation script which maps that to presentational MathML.
 >
 > I have been reliably informed that XSL does not allow specifications
 > that are expressive enough to do this job
since XSL does not exist, your informant clearly has Powers.

  > basically since it knows
  > nothing about maths, in the sense that it has no concept of arithmetic).
sounds like a computer reincarnation of me

 > > Would this not create similar portability/conversion/parsing problems
 > > that we have with TeX now if this were sufficiently powerful?
yes. but we'd be playing in the same swimming pool as the rest of the
 > > world

sebastian