Message-Id: <9303091752.AA07312@sc.zib-berlin.dbp.de>
Reply-To: Mailing list for the LaTeX3 project <LATEX-L@vm.urz.Uni-Heidelberg.de>
Date:         Tue, 9 Mar 93 18:51:38 CET
From: Frank Mittelbach <MITTELBACH@MZDMZA.ZDV.UNI-MAINZ.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@vm.urz.Uni-Heidelberg.de>
To: Multiple Recipients of <LATEX-L@vm.urz.Uni-Heidelberg.de>
Subject:      resend re fonts and LaTeX3
Status: R


From:	MZDMZA::MITTELBACH   "Frank Mittelbach"  8-MAR-1993 14:57:07.18
To:	TOLTX
CC:	MITTELBACH
Subj:	re: Fonts and LaTeX3


> Subj:	Fonts and LaTeX 3

Roffe said:

> First, just substituting one font for another (as with D. M. Jones' PS
> font package) isn't really good enough, because different x heights reqire
> different baselines. My point is, simply saying
> \family{somefont}\selectfont isn't good enough because the leading is
> still prepared for cmr.

this is certainly true, the flexibility of NFSS wasn't meant to stop
at this point, but in LaTeX209 the setting of baselineskip is
unfortunately quite hardwired into a style.

This certainly needs parameterizing, however I think that it would be
a wrong approach to bound values for leading to a font family. Whilere
there is probably something like a suggested leading for every family
it is nevertheless up to the designer to decide on the actual leading
for a given font and size. Also, it is difficult to imagine what
algorithm to use if two fonts are used on the same line (I really
can't believe that one would like to see the leading change there).

In other words I vote for settable leading in a more modular way, eg
without having to copy all such definitions for \large etc.

> Sometimes I'd like to keep cmtt (or, indeed, cmsf) and only replace cmr
> with a PS font. But this requires some work because font encoding between
> MF fonts and PS fonts normally differ slightly. So I'd like commands to
> switch between different font encodings.

NFSS release 2 which will go within the next weeks into beta testing
has the encoding scheme as one of its font attributes. However, the
correct solution here is to use standard encodings for all fonts in
the future, ie Cork encoding for latin based text fonts and other
encodings (which have to be agreed on, soon) for math, non-latin
languages etc.

> I'd like to be able to turn _off_ ligatures at large point sizes because
> ligatures look to cramped. But this can't be done from a macro
> unfortunately, you'd have to have to build tfm's for different point
> sizes.

yes this is an unfortunate optimization in the TeX program. But for
this very reason it *is not* an issue of LaTeX3. The only part which
concerns LaTeX3 is "can it be used within it?". Having such tfm files
you can use them with any macro package in TeX. And with NFSS (2) it
is fairly easy to use them in LaTeX.

> ....
> The command \AdobeGaramond would set appropriate baselines, correct
> mapping between input and output character encoding (to do this, it might
> have to read tfm files as well ... sigh).

as I said above I neither believe in global "correct" values for
leading nor in the idea that internal TeX macros are supposed to
correct deficiencies in the encoding of fonts. NFSS2 contains some
support for this, but in fact the encoding attribute is more or less
intended to be used only for change of scripts, eg if you think of
mixing cmr with wncyr fonts (latin and cyrillic fonts).

> Such a beast should be able to read not only PS fonts, but also HP
> softfonts I suppose, and be able to recognize Mac and Latin1 input
> encoding as well as the IBM codepages.
>
> Does this sound too far-fetched?

maybe nice to have, but I think much more important would be
additional encoding standards beside Cork and the use of the Cork
encoding as "the only encoding" for latin based fonts.

Frank