[pd@kubism.ku.dk: 8-bit fonts for TeX ?]

[mackay@cs.washington.edu (Pierre MacKay)]

You may be interested in a scheme I am almost ready to release,
which allows pretty free mapping of METAFONTS in the 128-255
area.  (down below too, but that can get you into difficulties
in the related TeX coding).  Why not have a hard and fast scheme?
Karl givess one reason, and the enclosed copy of a letter that
I never got around to mailing indicates another.

---------------------------------------------------------
  \opening{Dear Mr. Freytag,} 

  When I received the Unicode Draft over a year ago, I was working on
  other alphabets, and I am afraid I did not notice till today a curious
  omission from all the coding tables.  In a sense, it is
  understandable, since the missing letters are not part of the {\it
  official} alphabet of any country, but then neither are IPA
  characters, and they have been allowed spaces.

  I enclose proof sheets of the 11 characters I have recently been
  working on (lower case only, but they all have upper case siblings).
  These are used in the Roman-letter transcription of the principal
  Islamic (Arabic Script) languages, and probably have a wider
  international currency than many of the obscurer characters in the
  Latin 1 and Extended Latin set.  There are a couple of duplications in
  function, but no ambiguities unless ``s with bar under'' gets
  misused as a substitute for ``s with cedilla''.  Transcriptions concerned
  primarily with the rendering of Arabic Script in an Arabic context
  will use ``t with bar under'' and ``d with bar under'' where in a
  Persian or Turkish context, you are more likely to find ``s with bar
  under'' and ``z with bar under'' as representing {\bf 062B} Arabic
  Letter THEH and {\bf 0630} Arabic Letter THAL respectively.  There are
  literally millions of pages a year published making extensive use of
  the characters I enclose.  I am not sure when the equivalences were
  first agreed on, but it is well back in the 19th century.

  I became aware of the absence of these characters from Unicode when I
  was planning to remap a thoroughly idiosyncratic font table to some
  more general standard.  My first choice was Unicode, but I found to my
  surprise that half the characters I had developed have no identity
  whatsoever in any of the tables.  But every Latin letter periodical
  with a primary interest in the Near Eastern and Islamic studies uses
  these characters intensively, as do most of the monographs and books
  published in the field, and the Central Asiatic Republics in the
  Turkish language regions of the former Soviet Union, who speak with a
  somewhat different phonetic repertory than the Turks of the
  Ankara-based Republic are very likely to adopt some of this set.  I am
  a bit surprised that no one from the Research Libraries Group noted
  the extensive use of this character set in bibliographic references.

  I fear it is probably too late to do much about this now, which is
  rather a pity, since the particular design effort I have been working
  on is part of an attempt to develop a standardized text database using
  SGML.  But I felt it would do no harm to communicate with you about
  the matter, since I really do not know what stage the negotiations
  about Unicode have reached.  


-------------------------------------------------------------------------

The general idea is to have a distinct mapping file in which
accented characters are given names with purely descriptive
content, and to make that mapping file as easy as possible
to re-edit into an included TeX macro file.  It works
rather nicely, and by using an intermediate stage of
ligatures with a post-positive accent convention, I can
even feed several different ugly word-processor codings
through the same tex input file with only a change in the
small mapping file.  

If you're interested, I'll send a typical mapping file, and
in a short while I'll send the full schema.

Incidentally, it treats all of Don Knuth's computer modern
character descriptions (or any similar ones for that matter)
as read-only sources, to avoid the creation of dialect versions
of well-known characters.


Email concerned with UnixTeX distribution software should be sent primarily
to:	elisabet@max.u.washington.edu           Elizabeth Tachikawa
otherwise to:  mackay@cs.washington.edu		Pierre A. MacKay
Smail:  Northwest Computing Support Center	TUG Site Coordinator for
	Thomson Hall, Mail Stop DR-10		Unix-flavored TeX
	University of Washington
	Seattle, WA 98195
	(206) 543-6259