This document describes the procedure used to optimize font usage
for G-I.

1) determine the set of codes (glyphs) available in the font files
two "levels" of results:
 * list of codes _per_ font file (fontX.lst)

 * list of codes available _at least_ in one of the font files
(avail_codes.lst)

note that one particular codepoint can be in many font files.

 +---------------+      +---------------+
 | +-----------+ |      | +-----------+ |
 | | font1.ttf | |      | | font1.lst | |
 | +-----------+ |      | +-----------+ |
 |               |      |               |
 | +-----------+ |      | +-----------+ |
 | | font2.ttf | |      | | font2.lst | |
 | +-----------+ |  =>  | +-----------+ | => avail_codes.lst
 |               |      |               |
 |   ...         |      |   ...         |
 | 	         |      |               |
 | +-----------+ |      | +-----------+ |
 | | fontN.ttf | |      | | fontN.lst | |
 | +-----------+ |      | +-----------+ |
 +---------------+      +---------------+


2) determine the set of codes (glyphs) needed to display translation
files (.po)

 +---------------+      +---------------+
 | +-----------+ |      | +-----------+ |
 | | lang1.po  | |      | | lang1.lst | |
 | +-----------+ |      | +-----------+ |
 |               |      |               |
 | +-----------+ |      | +-----------+ |
 | | lang2.po  | |      | | lang2.lst | |
 | +-----------+ |  =>  | +-----------+ | => needed_codes.lst
 |               |      |               |
 |   ...         |      |   ...         |
 | 	         |      |               |
 | +-----------+ |      | +-----------+ |
 | | langN.po  | |      | | langN.lst | |
 | +-----------+ |      | +-----------+ |
 +---------------+      +---------------+


with the above information is possible to find out:

 ** missing glyphs :

    codes listed in "needed_codes.lst" and not contained in
    "avail_codes.lst" this information is associated to language files

    langX -> langX.mis
    langY -> langY.mis

 ** extra glyphs :

    avail_codes.lst - needed_codes.lst = xtra_codes.lst
    this information is associated to font files:

    fontX.ttf -> fontX.xtr
    fontY.ttf -> fontY.xtr

using the .xtr files is possible to remove unused glypph from the
associated font files; this does not cover the problem of overlapping
codes.