Quoth der Mouse:
> > clearly one of the best choices here, since it's the only standard
> > UNIX-compatible way to handle full Unicode range.
> 
> Well, the only one that uses 8-bit chars, maybe - I'd prefer to switch
> to 16-bit chars.  (Actually, wouldn't UTF-7 also qualify?  Yes, I think

I'm not sure utf7 was file-system-safe.  utf8 has the properties that no byte
of a multibyte sequence can ever be mistaken for another character, no matter
where you jump into the string, the ASCII range is identity mapped, and (it
follows) any NUL byte you see has to be a real NUL.

I seem to recall utf7 did not have those properties (actually it seems they
might not be simultaneously achievable in a 7-bit coding), though my memory
can be tricky.

-Chap