From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Mon Feb 03 2003 - 07:06:30 CST
In <8f7RLrBJcDD@3247.org> list-ietf-wg-apps-usefor@faerber.muc.de (Claus Färber) writes:
>Shmuel (Seymour J.) Metz <Shmuel+gen@patriot.net> schrieb/wrote:
>> In <H9L4rD.LzL@clw.cs.man.ac.uk>, on 01/31/2003
>> at 03:38 PM, "Charles Lindsey" <chl@clw.cs.man.ac.uk> said:
>>> I doubt anyone actually uses the Unicode Tag characters because, in
>>> bodies, there is the Content-Language header.
>> What about bodies that are primarily in one language with text
>> fragments in another?
>Well, it seems that this is handled by multipart/mixed.
Only by Turnpike that I know of (opinions differ as to how multipart/mixed
is supposed to behave in this regard, and implementations differ even
more). And, as a means of changing language (or even charset) in mid
stream, it is a huge KLUDGE.
But yes, that problem does exist in bodies but, since it is also a mail
problem, I don't want us to get bogged down within it.
However, a really ugly thought occurred to me last night. Suppose we
invent a new encoding for RFC 2047 (which is allowed), and also extend the
characters allowed in 'encoded-text' to include 8bit chars (but the
present encodings 'Q' and 'B' would never generate those). Then you could
say:
From: =?iso8859-1*de?Q8?Claus_Färber?= <......>
Of course, that encoding would only be allowed in Netnews, and you would
have to reecdoe it with Q-P at gateways.
But it would allow the chinese to continue to use their own character sets
in a manner that did not require any actual decoding by reading agents.
Does not solve the newsgroup-names problem, though.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5