Lately I've been wondering if it would be OK to see Unicode itself as
corresponding to MIME's "text/plain", and to see an extended Unicode
(i.e. language tagging and/or font tagging) as corresponding to
"richtext" (or one of its successors).
Erik, it should be noted for you that NO Japanese plain text is written
in Chinese Han.
How about Japanese names in Chinese newspapers? (In my view,
newspapers are "plain" text, and expensive books are "rich" text.)
(You probably wouldn't see full Japanese sentences printed in Chinese
Han, though, because Japanese uses Kana, which are not part of the
Chinese Han.)
Using Chinese Han for Japanese is like removing diacritical marks from
Latin.
How do you know what "removing diacritical marks from Latin" is like?
Even if you remove diacritical marks for diaeresis, umlaut, circumflex,
acute, ring and so on from Latin characters, the resulting text is still
readable
How do you know? (Are you fluent in French, or something?)
just as we Japanses can barely recongnize Japanese text
represented in unified Chinese Han, which often have extra dashes and
subtly different topologies.
I agree that there are differences, and I agree that they are subtle.
My *opinion* (which seems to be different from yours) is that *subtle*
differences can be adequately shown in "rich" text, and that Unicode
is good enough for "plain" text. And, as you know, there are some
Japanese that have an opinion similar to mine.
So it depends on how plain one feels "text/plain" should be, and how
rich richtext should be.
But MIME does not prohibit you from defining a "rich" charset for use
with text/plain. If you want to define a 10646-based charset enhanced
with language tags, then, by all means, please do so. Nobody is
trying to stop you.
Erik