Making the charset illustrations

Even if Unix has no full Unicode support yet, Unicode can already be used on Unix as a background encoding to manage various traditional charsets as it is done by GNU recode, the Linux console, the Lynx browser, the Kermit terminal program, or the Yudit editor.

Another instance proving this thesis for me, is my little set of scripts I wrote in May 1998 to maintain my charsets documentation: Two self-written scripts make it easy to generate a font and an illustration from a mere Unicode table instead of having to draw them manually.

For each charset, all I need is a textual Unicode mapping table defining the coded character set, for example iso8859-15.txt:

=20     U+0020  SPACE
=21     U+0021  EXCLAMATION MARK
=22     U+0022  QUOTATION MARK
...
=A3     U+00A3  POUND SIGN
=A4     U+20AC  EURO SIGN
=A5     U+00A5  YEN SIGN
...
=BC     U+0152  LATIN CAPITAL LIGATURE OE
=BD     U+0153  LATIN SMALL LIGATURE OE
=BE     U+0178  LATIN CAPITAL LETTER Y WITH DIAERESIS
...
=FE     U+00FE  LATIN SMALL LETTER THORN
=FF     U+00FF  LATIN SMALL LETTER Y WITH DIAERESIS

Unicode mapping tables have been freely available for many charsets and new ISO standards such as the parts of ISO-8859 revised in 1998 come with explicit UCS numbers.

Whenever new charsets are defined or errors in my tables are found, I only have to edit the textual definition. Then, with the unicoded mapping table I can simply say

make iso8859-15.{bdf,gif}

to automatically generate the matching bitmap font and an illustrative bitmap GIF showing the code chart:

This is accomplished through

a script called fixtable that brings the mapping table into my canonical form,
a script called bdfassemble to assemble the font,
a script called bdftogif to draw the image, and
a Makefile to put everything together

Beware that these scripts are poorly commented and still depend on my local file hierarchy and are still using an older glyph storage format that I originally had in mind for the GNU unifont with one file for each character.

This is another trivial example of how Unicode can be useful in Unix even though it is somewhat working behind the scenes.

Roman Czyborra
$Date: 1998/11/23 11:30:26 $