Yukihiro Matsumoto:

> Each string and regex object will be able to have information about
> its encodings.  Matching, indexing etc. will be based on that
> information.

   Are you planning on a string being a sequence of differently encoded
pieces or will each string use only one encoding? I've worked on software
that uses embedded markers when switching languages (which avoids problems
like matching words from different languages which are spelt the same such
as 'parole' in German and English) and it can be quite messy.

   Neil