> This assumes we're using Unicode for multilingual documents, and
> something like the ERCS for the parser, of course...
>
While Unicode solves the charset problem, negotiation between
server and client remains. For the time being there will not be
a client which handles all the languages represented by Unicode,
and so the issue arises what if a browser handles 3 of the 5 languages
used within a document. Will the server send the document? Is
it the browser's responsibility to deal with characters it cannot handle.
Is the current protocol for negotiation sufficient to deal with
these and similar cases?
-anoosh