Re: [Groff] ubuntu, groff and utf-8

[Michail Vidiassov]

Dear Alejandro and All:


as far as I understand, the situation with character composition
in groff is as follows

1) General case
 Opentype fonts may have character composition information,
  but it is hard to get out of the font.
 PS fonts used to have composition information, but it
  is of no use to groff, since it just described the precomposed
  accented glyphs, that can  be accessed by font reencoding.
 It is hard to make groff use the compositiion information from the font
  (requires a lot of coding or even redesign).
 There are few uses for composed characters, since most
  of the characters groff users need can be found precomposed
  in commercial or free fonts (*roff is not likely to be used to
  make a hebrew - russian dictionary with transcriptions nowdays).

 Result - no demand and the costs are high => do not do it.

2) Some fallback to print some more accented characters with
 base PS fonts - what Werner originaly liked to have.

 We are not the first to try to add more accents to base PS fonts
 (but, may be, we are the last).

 Solutions existed, such as:
 1) Instructions for adding additional characters to pfa files and
  describing them in AFM files.
 2) Tables of composition instructions for some of the base fonts +
   tools to make patched versions of those fonts.
 3) Automated tools, adding accented glyphs to any font
   (but in an ugly way, since only bounding boxes of glyphs are
    analysed in the best case).

  Since the tables from 2 were just (abused) composition sections of
  AFM files (CC lines), it may be easy to transform their contents
  into roff positioning commands (code for .fschar, what Werner wants)
  or a library of PostScript procedures.

3) Metrics for some decent PostScript or TrueType fonts are to
  be included in groff distribution.
  And, BTW, Carthago delenda est ;)
                           Sincerely, Michail
References:
1)
 Adding Composite Characters to Type 1 Fonts
  http://nl.ijs.si/gnusl/cee/type1.hacking.html

2)
 ogonkify:
  Coverage:
  http://www.pps.jussieu.fr/~jch/software/ogonkify/missing.html
  Home page:
  http://www.pps.jussieu.fr/~jch/software/ogonkify/
  Now part of:
  http://www.infres.enst.fr/~demaille/a2ps/
  Some additional accents:
  http://idea.hosting.lv/a/ogonkify/

 lout
  ftp://ftp.it.usyd.edu.au/jeff/lout/
3)
 a2ac
  Computes composite character shift information for any font
  and adds it to AFM file
  ( for later processing with afm2tfm  for use in TeX.
   afm2tfm reads and uses information from CC lines.)
  ftp://math.feld.cvut.cz/pub/olsak/a2ac
 cedilla
  Some kind of Unicode PostScript printer
  (no comments, since I can not read Lisp)
  http://www.pps.jussieu.fr/~jch/software/cedilla/