I wonder if it would be acceptable to punt MultiLingual messages to
RichText? In other words, a single MIME object would have a single
charset, (country/language) or whatever.
WAIT! While I think multilinguality should be achieved at plain text
level, I won't take the trouble to object your attempt. It is acceptable
to me that, with some code set such as ISO 10646, full
multi-european-lingual plain text is archieved and, still, rich text
is required for full multi-asian-lingual capability. We can just use
better code set, instead.
BUT, before jumping to the multilingual issues, we should settle
the issue of how *MONOLINGUAL* C/J/K or Devanagari messages can be
transmitted correctly as plain text, first.
With the following simple-minded specification:
charset=iso-10646
it is impossible to render monolingual Japanese text correctly
without external profiling information. Moreover, the external
profiling information is often unavailable.
Masataka Ohta