Let me just add some notes to the discussion of letterspacing in the
past few days:
I think that letterspacing is needed (although it is ugly) and so
should be available and easily usable within TeX, may be even if
this takes things up to the need for properly letterspaced fonts.
Let me give some examples, where letterspacing could be needed:
may be its main application is not within mathematics, but in the
humanities: Here you often have the problem ot markup many different
layers of a text, e.g. in a quotation when you have to show the
contributions of different previous authors or preliminary versions
or the like; here you easily run out of standard ways of markup.
Or even a more trivial example: sometimes you have to give a
quotation verbatim, copiying the original markup, if there was
letterspacing, you have to do that, and TeX should only help that
it doesn't look too bad.
I consider underlining text almost equally bad as a usual way of
markup (the situation seems somehow to be italics vs. letterspacing
and boldface vs. underlining), they both should be avoided whenever
possible, yet there are still some cases where it is useful, and
as TeX does make underlining possible, I see no reason to be so
restrictive vs. letterspacing (okay, it is more difficult to
do it properly, as special fonts seem to be needed)
Some comments on related topics mentioned:
As for Fraktur, Schwabacher came out of style as a means of emphasis
around 1800, afterwards only letterspacing was used (one can find
that somewhere in Tschichold's writings). So as the main usage of
Fraktur seems to render something historically correct, here letterspacing
is needed, too. I don't know about the treatment of ligatures, which
are rather freqent in Fraktur.
(I have seen a preliminary version of what is now a book about LaTeX by
Axel Heilmann, published by Springer (I can't remember the title), where
the author used yfrak for the main text because he likes it so much,
this rendered it almost unreadable and unusuable; so keeping the usage
to historic texts seems preferable...)
In mathematics, different kinds of emphasis could be achieved without
using letterspacing (e.g. italic for the body of theorems, bold extended
for the expression being defined, bold (at normal width) for things
like "Definition 3.1:" and sectioning etc.), so I don't see a need here,
even if some authors do use it (and I agree that Heuser is a well typeset
book in general).
Verbatim quotation isn't used here in the way mentioned above, so any
author could avoid letterspacing in any case.
Another thing related to letterspacing is Capitalization,
i.e. writing all-uppercase. Here, proper spacing is needed, too,
and TeX's kerning is correct for usual text and abbreviations,
but not for all-uppercase. I think Tschicholds "Meisterbuch der
Schrift" (there should be an English edition, too) gives some
examples of good and bad spacing, or it is elsewhere in his writings.
All-uppercase needs rather wide spacing, and TeX's lacking a good
way to accomplish this. May be a "titleing" font could be useful here.
The bad spacing seems to be the main reason why I dislike the
standard LaTeX headings (slanted and all-uppercase).
Johannes Kuester [log in to unmask]
Mathematisches Institut der
Technischen Universitaet Muenchen