>-- Other systems: Writing with a typewriter doesn't seem to be a serious
>alternative, and I doubt that it will be faster in most cases. And as
>for the other systems I can't find any point against TeX here other than
>"complicated to use" or "too time-comsuming". That is too commonplace
>as long as no better way is given explicitly. I simply doubt that other
>systems are superior here when one wishes to get the same quality as with
>TeX. So please give explicit examples, ...
...
>-- Rules for mathematical typography: Hans Abergs arguments here esentially
>boil down to "most mathematicians don't know about rules for typography,
>if they would know, they wouldn't like to follow them, and they couldn't
>care less about typography anyway". ...
...

  I just tried to indicate that some mathematicians do use others systems,
because they _feel_ it is better for them, knocking out some papers. But I
tried to carefully to avoid the question whether they are right or not, and
which mathematics they might work with, in order to not clutter this group
with lots of upset letters. :-)

  However, I think mathematicians in general are very sensitive to rules of
typography, and do want a nice, final typeset output (this was one reason
for not using the original LaTeX, because it did not conform to
mathematical standards, and was hard to reprogram). But the getting a
correct semantics comes in the first place.

  The LaTeX3 project is doing some very good work, and it would be good if
one could spend some extra effort on the math interface, so that
mathematicians would feel they do not need AMS-TeX or own non-LaTeX
macro-packages anymore, because LaTeX has so many other features, no single
person could develop.

  Concerning TeX, it is the ultimate in classical typesetting; by
"classical" I mean typsetting of kerned boxes of pictures (with no
knowledge of the graphics in the box). It seems me that the LaTeX3 now is
securing the principles of such classical typesetting, as far as TeX has
capacity; when this is done, the next step would probably be to find new
principles for typesetting.

  In fact, I am not sure that even the things that Frank Mittelbach wants
to do with LaTeX3, can be fully done without requesting a new TeX version.
This importance of the LaTeX3, namely, as an indicator on how TeX might
evolve, was already recognized by the NTS project several years ago.

>-- Commutative diagrams:
>This is surely rather hard to do in TeX. But that is because this
>is rather a picture than math text...

  Because of the limitation of TeX, diagonal arrows are hard to make, but
some things are surely possible. Michael Barr has a nice "diagram.tex"
package (on CTAN under his name) which can use digonal LaTeX arrows for
making commutative diagrams.

  In general, though, it is hard to reason that commutative diagrams "are
pictures", because they should normally have typeset information, and this
should come from the same macros used in the other formulas in the text.

  So this is really a flaw of TeX. Could one not integrate the METAFONT
drawing capacities into TeX? (But this is probably not a LaTeX3
discussion...)

  Hans Aberg