To: CyrTeX-en@vsu.ru, bnb@ams.org, vvv@vvv.vsu.ru
Subject: Re: ASCII-Cyrillic
Message-Id: <E13VruA-0005DE-00@topo.math.u-psud.fr>
From: Laurent Siebenmann <lcs@topo.math.u-psud.fr>
Date: Mon, 4 Sep 2000 10:47:02 +0200


Dear Barbara, Dear Vladimir,

    I was happy to finally see Barbara's unpublished
documentation corresponding to the full set of ligatures
in the wncy* fonts that Vladimir has kindly listed.  This
has long been a subject of puzzlement to me.

    Let's focus on a most important example:

101 cyre ; L zero cyryo ; L one cyrerev ; L two cyrie ;

This means that the four cyrillic e in the AMS encoding,
nowadays named:

     cyre   cyryo   cyrerev   cyrie 

could be typed

      e       e0        e1      e2 

for a Cyrillic TeX typescript destined for wcm* fonts.

    For the typist, this is just about ideal.  Currently
my ASCII-Cyrillic does not handle cyrie (round e) since it 
is not modern Russian, but the first three are typed:

      e       'o       'e

Thus ASCII-Cyrillic is equally good for Russian typing and
considerably better for reading. But it is less obviously
extensible beyond Russian...  By any measure, Barbara's
typing of four Cyrillic e-variants:   e     e0     e1
e2   is of considerable potential interest to typists.

     A point I would like to make here is that it is
useful and proper to distinguish the features/demerits of
a typing system from those of the typesetting system
situated downstream. What many of us need even today is
means of combining some acceptable 7-bit ASCII keyboarding
of Cyrillic with a form of typesetting that maintains
TeX's highest standards. And the AMS wncy* TeX setup
definitely does not --- see comprehensive reference below.

     I am expecting an active user of Omega to explain
how it can elegantly handle Barbara's sort of typing.

     But there are other candidates that operate happily
with today's TeX.  Briefly:-

 -- Emacs portable Cyrillic input keying for Latin/ASCII
keyboards (Vladimir?- this is news to me!)

 -- My recipe for a 8-bit Russian Macintosh-only software
Russian keyboard (already mentioned, I will explain more
later).

 -- My ASCII-Cyrillic typing scheme for ASCII streen fonts
with its ASCII to 8-bit conversion utility (using TeX).

     Infortunately, my TeX utility is not currently
programmed for Barbara's numerical suffix syntax
e1, e2 etc. And that might be awkward!

     In the TeXbook, Don Knuth, promised the ligature
performance that is still missing in TeX 3.  So Barbara,
you are not in any sense to blame for not forseeing these
deficiencies.  My guess is that the missing features were
crowded out of TeX 3 by 8-bit TeX. Knuth has gone to pains
to explain that (a) 8-bit TeX was not part of his original
vision of TeX, and that (b) in doing TeX 3 he felt pressed
for time by his other commitments. In these circumstances,
it seems to me somewhere between natural and inevitable to
abandon difficult ligature features whose ultimate purposes
8-bit TeX can easily achieve otherwise.  It has now become
Omega's pride and joy (I believe) to realize best possible
ligature features.  Such is my reading of history.

             Cheers

             Larry Siebenmann


PS. The best catalog of the extensive typographical
deficiencies of the "ligaturing" AMS wncy* system as used
under TeX 3 is in an article by

B. Jackowski and M. Ry\'cko, "Polishing TeX...", Euro-TeX
92, Sept 14--18, Prague CZ, pp 119-133.

This is out of print, I believe. Email to
B.Jackowski@GUST.ORG.PL for a PS version.

PPS.  The interest of the WNCY* fonts of the AMS (done by
Barbara Beeton and Tom Ridgeway) should similarly be
evaluated quite independently of keyboarding schemes, or
even typesetting systems.  Vladimir's support for the AMS
WLCY* virtual fonts is thus a step in the right direction.
Even today, the only freely available CM-compatible
Cyrillic fonts in Adobe Type 1 format are the AMS
Washington Cyrillic fonts.