Ok, that's what the font fallbacksystem is for. It's not done, but we should do it.If we start transliterating greek and cyrillic, we'll want to remove accents from latin alphabet too and stick to 7bit ASCII... I think that's a bad idea, we have utf-8 support and should use it.That is not a question of encoding, but lack of some codepages in the current font. I see no sense using font with complete codepages support for daily using, because it consumes lot of resource. AFAIR the all-in-one font from BitStream was about 15 MB big. And some fonts are good to render latin charsets, some to asian ones. By the way, the default font can be set on installation correspondent to user selected locale.
Ok, that makes sense. My choice would be to have both, then, and the transliteration be localized (so it's readable by who looks at it). Note ICU has a transliterator class which may be useful, if hte results are good enough : http://icu-project.org/apiref/icu4c/classTransliterator.htmlThink about some asiatic languages where the transliteration still uses a lot of accents and other strange stuff, and thus doesn't really help solve the problem.I prefer to have usability in mind. That are the names wich are used by those peoples to let the World to know about them. For example, I cannot pronounce Greek so I cannot to identify those guys somewhere else outside of our About System Window. But having transliterated names make theirs chance to become "my handshake and thanks" much bigger.
-- Adrien.
