[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: rsfs slant problems

Dear Matthias Clasen and friends,

 > There is only one thing for which we experiment with extended
 > skewchar trickery: underaccent skew. It seems natural to use
 > the `free side' of the skewchar to encode this information in
 > kerning pairs.

Seems perfectly good subject matter for e-TeX and a sound
epsilonic approach.  Similarly, a single new skew
character suffices to adjust (sub- and super-) script
positions as a function of the character adjoined.  (Add
still another skew character for pre-scripts if that seems
important, which I doubt.)

 > If more shape information has to be associated with glyphs, we
 > should think about a redesign of the tfm format. It would be
 > good if this could be coordinated with omegas ofm extensions.

A sound NTS-Omega approach.  

The philosophies and the code of e-TeX and NTS are radically 
different.  The projects should be independent.

The notion that the e-TeX fish can metamorphose into the 
NTS fowl (bird of Paradise) seems strange indeed to me.

 > I don't think that the `skewchar trickery' should 
 > be extended too far, since it eats up valuable slots

In view of some of the rubbish going into the new math
encodings, there should be place for a few skew characters ---
the more so as they can moonlight as accents, a la Knuth.

I have never understood why TUG fonts are, like CM, on
principle, cram-full, thereby preventing upward compatible

In the CM case, it was (yes?) because Knuth did not trust you
to assure worthy evolution of his CM fonts.  Since you surely
trust yourselves, why not leave many characters undefined and
add more characters, say once a decade? Time flies: if this
policy had been followed with the Cork 1989, norm, it would
already be time to debate an upward compatible revision.


                    Larry Siebenmann