At 13:44 97-04-14, Frank Mittelbach wrote:
>Hans Aberg writes:
> > My comment was not intended as polemics. As a matter of fact, several
> > mathematicians I know refuse to use LaTeX, because it cannot provide the
> > mathematical typesetting they think are needed. The situation improved with
>and what are they using if i may ask? something else than TeX? or plain TeX?
There are all kinds of variations, depending on the work field. Of the
TeX users, the more extreme would write their own macro packages, perhaps
just including the symbols from AMS-fonts that they need.
Others still use AMS-TeX.
The thing is that LaTeX departs too much from what you would expect in a
"normal mathematical paper" (but I do not claim there is a universal
agreement of what a "normal mathematical paper" should look like). (One
example might be the equation tags story.)
In addition there are people that use other things than TeX. One
variation is using word processors and equation processors, which they do,
because they feel TeX is too complicated, and they cannot write
sufficiently fast using TeX. One requirement for this to work is that you
work in a field which is not too requiring from the typesetting point of
Perhaps there are people at IAS (Institute for Advanced Study, where
Einstein had a job), and such places, still using a secretary and
typewriter (but the secretaries probably use TeX these days). One example
of how such math might look is SGA, Grothendieck's Algebraic Geometry
Seminars, published in the Springer Verlag Lecture Notes in Math series,
and this series should have many other typewritten math manuscripts.
So there are all kinds of variations.
Frank Mittelbach wrote:
>(and yes kerning is a problem but that was looked at as well and
>within TeX you are limited to kern only within one font so there is
>unfortunately no way to get *all* kerns you ideally want. for this we
>have to wait for omega)
If you try to write a complicated mathematical paper, which is actually
going to be (re-)typeset, then the problem is that each correction may
generate new errors, and subsequent corrections may not be able to remove
all known errors. So, at some point you have to give up, accepting that not
all errors can be corrected.
So there are two problems:
How the author should be able to quickly turn out a manuscript with
typography that is semantically correct, and how make a good looking
Therefore, if you just want to knock out a manuscript, getting all the
kerns right is not a major concern for the author, but it would be a big
concern for the typesetter. :-)
So, returning to LaTeX3, it would help making some better user
interfaces, making the AMS-LaTeX package and LaTeX2e interface better,
making sure that there is a good arrows package, etc. It is less important
if the stuff ends up in a core distribution, or in a standard package, such
as AMS-LaTeX, but it should not be in LaTeX3 independent packages.