Dave,
if the disagreement had been about a content-subtype, I would agree
with you wholeheartedly.
BUT: The current text of the document is:
"Implementors are discouraged from defining new character sets for mail use
unless absolutely necesary".
(section 7.1.1, page 22 in PostScript)
The rest of the text strongly supports the reading that implementors should
AVOID other character sets than US-ASCII, ISO-8859-X, and ISO-2022-JP.
This is direct conflict with the opinion held by the NETF and EUNET
meetings that considered this matter, where MNEMONIC was preferred.
MNEMONIC also has a large user base.
The result is that when people from the NETF/EUNET community come to
American vendors (or others that regard RFCs as "holy writ"), and ask
for MNEMONIC support in their mail UAs, they will have problems.
So, I believe that the "false beliefe that failure
to have citations included will cripple MIME's utility" will, indeed,
cripple the utility of MIME for those people who have decided that their
preferred character set is MNEMONIC.
In short, I find it very hard to say that there is consensus on the omission
of MNEMONIC from MIME; in Santa Fe, I thought we had consensus on the
inclusion of MNEMONIC in MIME.
Harald Tveit Alvestrand