Message-Id: <9311042323.AA28475@sc.ZIB-Berlin.DE>
Reply-To: Mailing list for the LaTeX3 project <LATEX-L@vm.urz.Uni-Heidelberg.de>
Date:         Fri, 5 Nov 93 00:22:14 +0100
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:      Re: latex2e, nfss2 and internal font selection commands
Status: R


> Subj:	latex2e, nfss2 and internal font selection commands

> Our users often use pictex.tex and the pictex-exports of the program 'xfig'.
> The pictex.tex-macro only uses the internal font selection command \fiverm,
> but 'xfig' exports code, which contains many commands as \tenrm, \sevenrm,
> etc (mostly of no need, but it does). I see, that it is possible to write
> such things as
> \newcommand{\fiverm}{\fontfamily{cmr}\fontseries{m}\fontshape{n}%
> \fontsize{5}{7}\selectfont}
> in pictex.tex (though there should be no changes in the original file),
> but I do not know, how to change 'xfig', so that it does not export
> these internal font selection commands. So I would have to provide
> a style-file, which defines these commands (in the way like in the exmple
> above) or dump them in the format-file. But as I understand, the latter
> is not a possibility, when all formats of latex2e should be the same (we
> want to decrease the number of different formats, don't we?).
>
> Frank explained, that the internal commands are different at various sites, so
> it is understandable, that latex2e will not support them.
> So the question remains: What is the recommended way to get rid of such
> internal commands, when they are produced by some software (or: who updates
> the styles (such as pictex.tex) to be nfss2-compatible). Global definiton or
> style-files do not seem to be a good solution, do they?

we are aware of this problem. First of all, the use of \selectfont is
not an optimal solution to implementing something like \fiverm (by the
way good old lfonts doesn't have fiverm :-) for applications like
pictex where \fivrm is essentially used to type just a single
character, because there is no need to set up the whole machinery of
math typesetting for this.  Second, you do not ncessarily want to bind
something like \fivrm to cmr fonts (if you use times for example).

For this nfss2 under LaTeX2e will offer a better solution which is:

\DeclareFixedFont{\fivrm}{OT1}{cmr}{m}{n}{5pt}

this will define \fivrm to be really the same as fivrm in lfonts
namely

\font\fivrm= cmr5

but using standard nfss concepts and thus being fully transparent to
layout changes (eg if on another side the .fd file for OT1cmr contains
other fonts, for example, scalable ps versions of cmr it will still
work properly.

The compatibility mode will most certainly define the few ``standard
misuses out of lfonts.tex' so that something like pictex is accounted
for. However, programs that hardwired output code not being officially
latex should probably be changed in the long run to use a proper
interface. In case this doesn't happen for, say xfig, i don't think it
is much of a problem to provide a style xfig that defines the
functions written out by this program in a way that the user can run
it under latex2e without having to miss the benefits of the new
functionality because the program is running in compatibility mode.


> What do you think about this problem? I think in the compability-mode
> latex2e will not be fully compatible with latex209 :-).

There is no way to make a program that allows user extensions fully
compatible when adding anything since this already might hit a user
definition. However, when i said that using internals of lfonts was
never portable i meant just this. For example, on the installation in
Mainz we had \tentt and \sevtt preloaded because they were used in
math. However, on other installations such a font had probably a name
like \\ptt\@xpt, not a very typable thing with two backslashes in
front. Something like this was okay as long as nobody tries to send
documents containing \tentt around.


> So I fear that
> some people will not upgrade to latex2e because of this. I hope I am
> wrong. Maybe, when a good solution is provided.

I don't think that there are many of such instances so that the
compatibility mode should not be a problem. As for upgrading, using
pictex in latex requires certain startup files and i don't think it
should be difficult to correct them.

LaTeX2e is a bit different from installing, say, AmS-LaTeX on a
certain site. It is not another LaTeX flavor, it is the official
LaTeX. And it will be the only one supported from the point it is
released. Thus, if something new is claiming to run with LaTeX it has
to run with LaTeX2e in the future. We will try very hard to have the
largest possible compatibility (that is theoretically possible) with
user documents (ie 102% :-) but when there had been changes to
internal macros in the past (due to bug fixes, for example) packages
that made use of such undocumented features had to upgrade to run with
LaTeX properly, and they did so far. I'm confident that this will
happen in thefew instances where it is necessary.

It is claimed that there are sites that still use amr fonts,
fortunately, in most cases people from such sites can nevertheless
send you their documents and your latex will typeset them. But if they
use a style that says \font\foo=amr10 then most of you will probably
have to correct it because it isn't something support by your latex,
is it?

good night

frank