>  > http://www.Uni-Mainz.DE/cgi-bin/ltxbugs2html?pr=babel/3046
> Johannes talkes about the difficulties in providing this, but is it really
> something one wants?
>  - except when using mule (or emacs) one doesn't (automatically) change input
>    encodings when
>    changing a language in the middle of the document.

I have tried so, using emacs (for an English-German-Russian
doc), and was not quite happy about it.  Then I found out that a utf-8
based encoding is work-in-progress, and this would be A Good Thing to
pursue and integrate into the kernel in mid-term. See the current
version at
        http://www.unruh.de/DniQ/latex/unicode/ ,
CTAN is not up-to-date :-(hark, Robin).

Its author Dominique Unruh (cc:) also put some effort in an
experimental automatic font encoding selector (unsupported/autofe.sty) which
is used primarily for use with CJK (afai understood, I don't speak any
fareast tongue).

I would not vote for following the Java train to all-unicode-processing now,
but it would be fine if latex entered the platform.

Karsten Tinnefeld                       [log in to unmask]
Fachbereich Informatik, Lehrstuhl 2                   T +49 231 755-4737
Universitšt Dortmund, D-44221 Dortmund, Deutschland   F +49 231 755-2047