[tex-fonts] Re: [Fontinst] On the proper look of the \AA

Thierry Bouche thierry.bouche at ujf-grenoble.fr
Tue Jan 20 09:26:20 CET 2004

Just to add entropy...

I would expect the ring to touch the A in the case of angstrom, and to
behave like any other accent in the case of the swedish letter. I can't
explain why! Given the list of glyphs available in legacy CM, I would
expect that \AA represent an angstrom and not a swedish letter...

LH> Aha. I was rather thinking about what a font designer would do.

LH> Actually, as a general rule I would think that _if_ the author of a new
LH> font encoding includes some trick like the

LH> \DeclareTextCompositeCommand{\r}{OT1}{A}
LH>    {\leavevmode\setbox\z@\hbox{h}\dimen@\ht\z@\advance\dimen at -1ex%
LH>     \rlap{\raise.67\dimen@\hbox{\char23}}A}

LH> of OT1, then he should design this trick so that it can work for arbitrary
LH> font designs.

this is nonsense, there are no universal rules like that. You would
expect an Ultrablack font to have strong touching of A and ring, which
could happen to be a gap in the Ultralight version of the same font.
Typically, accents over capitals in good fonts are relatively lower than
over lower case, but they also are designed differently (have a look at
Utopia, e.g.).

Another place of latex that does pseudoglyphs that behave relatively
well with CM but usually wrong with other fonts is \textsuperscript.
Comparing built-in superior letters with \textsupercript shows that
you can't compute the size and placement of typographer's superior
letters from ex, em, capheight of a given font (although a typical way
of doing it could be to use \footnotesize rather than \scriptsize and to
move it so that a superior x is a hair below the top of an X).

However, this is not really important, imho: what is expected from an
encoding file is just to declare glyphs, and maybe to provide default
substitutions. What is expected from font support is a style file which
overrides these defaults. But, of course, there is a problem when you
load two style files, say one for a sans and one for a serif, where
macros could be redefined according to each font idiosyncrasies.

 Thierry Bouche                                      

More information about the fontinst mailing list