Roozbeh Pournader writes:
 > Sebastian Rahtz wrote:
 >
 > > these days we'd have it retyped :-}
 > So what about authors' full control of line/page breaks? Forget it?
absolutely. how can you control the line breaks when you dont use the
same font as me? you want us to insist that you buy commercial fonts?

 > > cos the AMS use TeX to typeset. why dont Elsevier's typesetters?
 > What is the Elsevier's typesetting engine?
 >
you misunderstand. we dont typeset. we send articles off to
typesetting companies around the world. they use whatever they want.

 > > oh, i agree. good LaTeX is very good. no argument.
 > And I can't forget the really good documentation in TeXLive CD
 > in many formats which I could not believe. It seems so
 > clean (well, except for some little cases) so that I thought it has
 > been prepared from scratch for each format... (don't believe :).
no, it all came from a Makefile, no tweaking at all. but that took me
quite a while to set up, and enforces constraints on the markup. each
time I do such things, I have to tweak the setup or the document a
little - seldom is the work flow of production quality

re XML
 > to this list. But my viewpoint is different. It is good, and has a good
 > future, but it will be misused by Microsoft or other companies, and
 > come into non-portablity problems (just like HTML). I think LaTeX's
I dont entirely agree. its hard to break XML itself, even for Bill
G. what he can (and will) break is the technology that makes XML
useable, like XSL.

 > Having sebastian's comments in mind, XML is taking the world,
 > so why shouldn't we support XML ideas as soon as possible? Why not
 > make LaTeX3 as robust, as simple, and as powerful as possible?
 > Even as a XML publishing engine, LaTeX has a good future,
 > and even if XML does not take the world, LaTeX is far from
 > perfect. I think we must discuss more exact problems (like
 > that frontmatter idea, table formatting problems, etc.).
hmm. i see the problem as the perceived differentiation in XML
Paradise between transformation and formatting. LaTeX has no
transformation layer, so its relegated to the formatting end; but by
that time half the interesting work has been done (eg TOCs), so LaTeX
becomes over heavy

 > But I have seen that no academician cen resist not-using TeX, except those
 > who don't write.
how many do you want me to show you before you believe that there are
eg computer science and physics academics who never use TeX?

 > Yes, yes. Dear Sebastian, this is very important. LaTeX has got
 > somewhere in between, between visual markup and and a logical one.
 > How do you think a math author trying to introduce a new notation
 > can use generic MathML?

macros, you mean? you want to say, "from now on, we will write this as
\Dipsy{34}"?

conceptually, the same way you do at present - define a new markup
language, with an element <Dipsy>, and write in it. then write a
transformation specification for how that maps to low-level
MathML. keep that separate from the typesetting/style specification.

but this enters deep waters and begs many questions. its equally
possible we'll be talking about special purpose DTDs with multiple
inheritance from things like MathML and CML, or an extensible
MathML. I suspect David Carlisle can give a more realistic "future
view" that me. but *conceptually*, its no problem - write any language
you like. what the user interface will look like is less clear. what I
do claim is that IF you want to talk about <Dipsy> elements, it is
reasonable to ask you to spend a little time defining the semantics of
Dipsy in an abstract way, not jumping straight into
\def\Dipsy#1:#2\\{$\expandafter\expandafter\expandafter\takeawaythenumberyoufirtstthoughtof\futurelet$}

Sebastian