> I'm about to make a fresh conversion of the TEI-L archives (for import
> in MarkMail this time). Ideally, I would take this opportunity to
> improve on the issues David raised about character encoding in the
> nabble mirror
> (http://tei-l.970651.n3.nabble.com/TEI-L-mirror-s-tp1878767p1878945.html).

It's tempting to suggest probability-based langauge / character set dowsing a la TextCat http://www.let.rug.nl/~vannoord/TextCat/

cheers
stuart