[Fontinst] A bug in fontist 1.927?

[Lars Hellström]

At 15.55 +0100 2005-01-22, Peter Dyballa wrote:
>Hello!

Hmm..
Subject: [Fontinst] A bug in fontist 1.927?
is getting a bit old. Perhaps you should pick one that more accurately
reflects then subject at hand?

>Trying to integrate TrueType fonts into TeX by converting them to
>PostScript via ttf2pt1 and fontinst I use fontsmpl.tex as a means to
>check my success. Now it happens that I see for monospaced fonts in OT1
>this for \l and \L and unusual settings for \textendash and \textemdash
>(Lucida Sans Typewriter @ 200%):

This is, although buggy, perfectly normal. What LaTeX calls OT1 is, as
mentioned in the fontinst manual, really five different encodings (all of
which were cooked up by Knuth) and thus some commands fail for some fonts.
This is yet another reason T1 is preferable to OT1 as text encoding.

The differences are formally spelled out in
fontinst/doc/encspecs/ot1draft.etx.

>I see the same when using a PostScript font, Adobe Courier. Here is
>another excerpt that shows that the \. and \H accents are wrong:

Again, this bug is documented.

>And there is one more thing with monospaced fonts, the shape of quotes!
>Here is an example from Monotype's TrueType font Courier New T1
>encoded:
>
>mcr-t1.png
>
>and this is the corresponding piece of fontsmpl.tex OT1 encoded:
>
>mcr-ot1.png

Yes, more of the same problem.

>Is TeX making mistakes? Is fontsmpl.tex incorrect?

Yes, it could be argued that fontsmpl.tex is incorrect, as much of what it
will show is "broken by design", but OTOH I don't think it has any special
provisions for OT1 at all. (fontsmpl.tex just goes ahead and tests a bunch
of accents and characters that aren't in that font.)

>Or has fontinst
>1.927 a few bugs? Why is \textquotedbl empty? Isn't it the glyph in
>slot '042/"22? \textquotedblright does not exist. Or am I
>misunderstanding TeX's use of double quotes?

These things are the way they are for compatibility with the Computer
Modern fonts. Had it been done as an orthodox interpretation of ot1enc.def
would have required, then very many documents written using the CM fonts
would have been wrong if set with other fonts. In the early nineties, the
CM fonts weighted much more heavily as a standard than did LaTeX2e's NFSS.

Lars Hellström