> Tomas, I think that you're confusing the intent of UTF-8 with UTF-7. > Can you explain where helps UTF-8 in the communication through 8-bit > channels? You can also send UTF-32 (or any other encoding) over 8-bit > channels, chopping each char in 4 bytes. UTF-8 strings can be passed through dumb processes based on the 8-bit C library; UCS-2 strings, for instance, cannot, because any leading nulls will be interpreted as string terminators; that's all I meant (I suspect this is responsible for the proliferation of utf-8 in the *nix world, because it allows to start using Unicode without having to rewrite the entire system).
Tomas
