"MIME: Message Header Extensions for Non-Ascii Text"
(Maybe the docs should say "MIME: Part 1" and "MIME: part 2"?)
I think that "MIME Part 2: Message Header Extensions..." would go a
*long* way to helping people read these documents together. We have
"implementation experience" from the "proposed" period that they don't
tend to do so.
2. Set of allowable charsets:
Suggested fix: allow any of the charsets defined in RFC1341bis for
use with the text/plain content-type, or any charset registered
with IANA for use with the text/plain content-type, or any extension
charset name beginning with "X-".
Keith, what we know from years of experience is that we are much
more likely to be fouled up by some unpleasant artifact of a half-baked
822 implementation than by anything that implementation thinks of as
"body". I suggest that we try to be very narrow and conservative about
this, becoming incrementally more permissive only as the need is
demonstrated. That very conservative model would be:
-- No permission for X-character sets in 1342bis
-- A requirement for standards-track RFCs, not just IANA
registration. The practical difference is that the latter permits a
systematic community review of potential interoperability problems and
adequacy of definition. IANA registration, as described in RFC1341,
does not imply either of these things unless IANA applies its
considerable discretion and imposed additional procedures or review.
--john