Rick Troth answered to John G. Myers:
If there is going to be some restriction on 8-bit characters in news,
be it "no 8bit characters in headers" or "no 8bit characters at all",
...
I think "no 8bit characters in headers" is a reasonable requirement
even if the (majority of) the transfer path is 8-bit clean. Headers are
(today) more for the software than for the user. It's nice that they
are humanly readable (I STRONGLY support that everywhere) but headers
don't have to support full ISO-8859-*.
It is true that headers are MOSTLY for the software, but
a set which humans do read a lot are:
Subject: (whole content)
From: (names)
To: (names)
Cc: (names)
Keywords: ...
My personal name is devoid of any 8-bit characters, however some 10%
or so (a wild guess based on a peek of phonebook..) of Finns have names
which use chars: Ä Ö (and Å) (TeX: \"A \"O \AA )
And of course we write text which uses those characters a lot, and looks
silly without the diacritics...
For the News (and news alone) I would suggest a rule:
- For a C-T: TEXT/PLAIN message in C-T-E: 8BIT form the defined
charset can be applied also for the headers UNLESS they use
RFC 1522 encoding to define othervice.
Such rule would solve things for the MOST CASES of news usage, though
naturally NEWS <-> EMAIL relaying systems need to be smart and do
proper MIME conversions..
Especially the sender implementation can be done with a mimimum impact:
Simply add those three header lines.
Like Keith has commented earlier, "Just-send-8-bit without labeling
the character set is foolish"
Not that Masataka's ISO-2022.. proposal would be that bad, it just
feels to me a lot more like an interface in between MIME-parsing system
(pine, for example), and the display terminal..
--
_.John G. Myers Internet: jgm+(_at_)CMU(_dot_)EDU
LoseNet: ...!seismo!ihnp4!wiscvm.wisc.edu!give!up
/Matti Aarnio <mea(_at_)nic(_dot_)funet(_dot_)fi>