Re: psnfss and lw35nfss

[Constantin Kahn <kahn@math.uni-hannover.de>]

Sebastian Rahtz writes:
>Alan Jeffrey writes:
> >  >  b) are the kermings in 8r.mtx bad, as Thierry suggests
> >
> > I reckon so.  The kernings in latin.mtx are the best list I could
> > manage...
>why has no-one said this for 2 years? i see it has Constantin's name
>on....

Sebastian,

Could it be that no one is using 8r-fonts? Could this be related to the
fact that the metrics on CTAN were essentially useless for 8r-typesetting
since at least February last year (when I first told you about inconsistent
character dimensions and LIGTABLE problems in psfonts)?

And, yes, 8r.mtx indeed has my name on it. You may remember that the
purpose of this file (along with some internal changes to fontinst) was to
create 8r-fonts with fontinst rather than with afm2tfm (or afmtotfm? I mean
the one that's associated with dvips, just can't remember which) which was
your first approach. To this end I essentially created a stripped-down
version of latin.mtx sufficient to do fonts in 8r encoding, plus kerning
pairs for lower-case accented glyphs analogous to those for upper case
glyphs. This addition of the lower-case kerning pairs was done to mimic the
results obtained with afm2tfm if memory serves me right (and it also seemed
more natural).

However, I've just run tftopl on your new ptmr8r.tfm file. The resulting
ptmr8r.pl doesn't contain a LIGTABLE at all! So there are *no* kern pairs
(and of course no ligatures which was observed before). Funnier yet, there
aren't even FONTDIMENs in this tfm. (Didn't know this was possible...)
[Note: I checked this with DirectTeX and with OzMF on a Macintosh; my own
Unix box presently is a shameful mess and not up to the task, so you may
wish to check for yourself.]

(Incidentally, the only 8r-psnfss-metrics available for downloading these
days which actually incorporate the kerning pairs/ligatures from 8r.mtx
seem to be the ports to Textures' proprietary metrics format which I've put
on CTAN. And that's just because I've never updated them to the screwed up
psfonts released early last year... ;-)

Sebastian, before jumping to conclusions and messing with the 8r.mtx file
(which may not be the culprit after all) can you *please* (double-)check
your scripts/methods? If at all possible then please reissue correct
8r-metrics which are *truly* based on 8r.mtx (which the current ones in
psfonts-beta definitely aren't), so that people can verify their claims
about deficiencies in the 8r kerning pairs. Thanks.

(Otherwise, I would not object to any decision to remove the lower-case
kerning pairs from 8r.mtx if they really cause visual quality problems.)

Constantin

--
Constantin Kahn, Am Fuhrenkampe 80, 30419 Hannover, Germany
Phone: (+49) 511-762-5380 or (+49) 511-758528 (home)
Email: kahn@math.uni-hannover.de