On Wednesday 09 January 2008, Eddy Petrișor wrote: > > Also, this is a completely separate issue from what we were discussing > > as it would _still_ result in duplication of latin characters. > > No, since the range in question wouldn't be pulled in by anything (the > codepoints would belong to the language or the unicode range that > obviously belongs to the language using the font). > > For instance for the Amharic language, the font would end up having > only the glyphs for the codepoints which are used in the translation > and all the Ethiopian unicode points. Anything else falls out of the > font. But most translations, even in other scripts, still use quite a lot of latin glyphs. You are also forgetting that quite a few strings with latin glyphs are generated *at run time* (e.g. udeb names and descriptions in anna; file systems and mount points; names of LVM volumes; time zones; ...) and you really do not want the situation where some latin glyphs are taken from the font from Amharic and some other latin glyphs are taken from DejaVu or FreeSans.
Attachment:
signature.asc
Description: This is a digitally signed message part.