🚀 go-pugleaf

> if a message was tagged (say as being latin1, or shift-JIS, or big5 or
> 8859-9), we have no problem

The problem is that, currently, not very many messages are tagged.
Hopefully, MIME tagging will become more widely used soon.  But there
are other ways of determining the character sets used in messages.
For example, the Japanese have been using an encoding now often
referred to as "ISO-2022-JP", that contains escape sequences that
indicate the character set of the following bytes.


> question 1 is: is this a plausible tactic, and is there a standardised
> list of character sets that are likely to be supported (or even
> better, must be supported)?

That depends on what you actually want to do.  For example, if one of
the researchers at your lab can read Japanese and wants to subscribe
to a mailing list in Japan, then that person should have ISO-2022-JP
software installed.  On the other hand, if someone wants to read the
Japanese newsgroups (such as fj.*), then you need to arrange a
newsfeed from some site that carries fj.* and whose upstream sites
don't strip ESCAPE.


> question 2: given a plausible answer to question 1 (say, MIME), would
> someone with such a system volunteer to conduct some tests of it with
> me?

I can send you ISO-2022-JP mail, with or without MIME headers.  I can
introduce you to some Japanese mailing lists.  I don't actually know
how to get fj.* to you, but I'm sure I can find out, if needed.


> question 3: how do i go about registering a sensible encoding (like
> the FSS-UTF mentioned above) so others can use it?

MIME charsets are registered using the procedure described in MIME
itself.  You may want to circulate a draft before registering the
document.
-
-- 
Erik M. van der Poel                               erik@poel.juice.or.jp