Martin,

> >  Those questions affect the string/character encoding, and my
> >conclusion is that the only thing should go there is the language
> >info.
> >
> 
> Do you mean that you want to differentiate an american english $a from 
> a german $a or even from a british english $a ?

  Not unless you really want to do so.  Latin-1 languages will share
the same language info and the same language tag value (which is
zero).

  It would be convenient to have if you want to do #asUppercase and
#asLowercase conversion properly for a latin-1 multilingual text,
though.

> I hope I don't 
> understand you correctly here. What kind of "language info" should go 
> into the character encoding?

  Language info for the Chinese characters, mainly.  If you're
interested in it, googling "unicode bone problem" and looking at the
FAQ section of Unicode.org may be interesting.

-- Yoshiki