#> > ISO 2022 raises several additional problems, arising primary because
#> > it is not a character set but a collection of separable rules. If we
#> > can avoid providing for it in headers, we can avoid replaying all of the
#> > arguments about it that ran through the list in the (northern
#> > hemisphere) spring.
#An email software developer that tries to sell a product in Japan that
#does not support the JUNET profile of ISO 2022 is just being
#incredibly stupid.
#...
#Since email software that handles Japanese is going to have to deal
#with their 2022 profile anyway, I would think that it is only natural
#to provide for the use of 2022 in the headers, obviously through
#further encoding and/or quoting to avoid the special 822 characters.
Ok, to replay one of the arguments/discussions from last spring...
The key difference between what I was trying to say and what Erik has
said here is the use, in Erik's text, of the word "profile". If one has
a non-ascii-in-headers model that can include an ESC character, the
JUNET ISO 2022 profile is no more difficult to handle than any other
character set. Indeed, because it is intrinsically 7 bit, it is a
little easier than some.
I have no problem in principle with calling out any particular 2022
profile, permitting it, and designating it with whatever mechanism is
used to designate other header character sets.
The problem is only with identifying "ISO 2022", without indication of,
and prior agreement on, a profile, as if it were a character set.
--john
p.s.: I would caution, however, that there is mail software (mostly
receiving UAs) on the network that will strip apparent escape sequences
(or the ESC character) on the theory that they are protecting users
against screen-killers. JUNET profile 2022 with the escapes removed is
roughly as information-damaged as the ISO8859 non-Latin sets are by
bit-stripping as an 8->7 conversion process. One of the important
things about JUNET profile 2022 is that it is rarely seen outside the
JUNET community. I have no idea what explicitly authorizing it in an
RFC would ultimately do to the rest of the nework.