RE: 1 in 1000

From: Michael Everson (
Date: Wed Apr 26 2000 - 09:34:27 EDT

Ar 14:41 -0400 2000-04-25, scríobh Robert A. Rosenberg:
>At 02:02 AM 04/25/2000 -0800, Michael Everson wrote:
>>Unicode doesn't model languages. It represents
>>writing systems, and while imperfect, the model does a very good job.
>Unless, as a case of an imperfection in the representation, you are writing
>Serbian where there are a few characters that need codepoints different
>from the Russian (and other Cyrillic) values so they can be distinguished
>from the Russian characters but the committee refuses to issue them since
>"They are the 'same' letters as the Russian/Cyrillic ones so you can use
>those codepoints".

But they are only distinguishable in some italic fonts, and that kind of
font distinction is not a plain-text character distinction.

>If you do not remember this issue it has to do with the fact that the
>shapes of certain Serbian letters are different from their Russian
>Counterparts when using an Italic Font.

Of course I remember the issue. Please read the Technical Report on the
Character/Glyph model.

>By assigning separate codepoints
>for these characters (and having the keyboards emit the correct codepoint
>depending on if you are using a Russian or Serbian mapping), the correct
>glyph will thus be displayable without needing to swap glyphs based on
>language (they map to the common glyph in non-italic fonts while having
>separate glyphs in the italic fonts).

And the user will never know unless he or she is using an italic font?

Michael Everson ** Everson Gunn Teoranta **
15 Port Chaeimhghein Íochtarach; Baile Átha Cliath 2; Éire/Ireland
Vox +353 1 478 2597 ** Fax +353 1 478 2597 ** Mob +353 86 807 9169
27 Páirc an Fhéithlinn; Baile an Bhóthair; Co. Átha Cliath; Éire

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:02 EDT