[EAI] UTF-8 in Message-IDs

[Frank Ellermann <[e-mail]>]

Hi,

looking for something else (= state of the art wrt EAI
and SPF) yesterday I stumbled over the "81 IETF EAI
meeting minutes" in the mailing list archive, with a
note about "UTF-8 in Message-IDs".

That was certainly (bad) news from my POV, and whining
about it off list I got the tip that waiting for an
IETF Last Call is a bad plan when this could be also
addressed before WG Last Call.

>From my POV the "globally unique forever Message-ID"
in mail and news is in essence an early form of what
is now a GUID/UUID in other protocols.  Maybe a bit
more so, in theory there could be UUID collisions,
and that's not possible for Message-IDs (in theory).

This theory is relevant for NetNews, gateways from or
to NetNews, mail and news threading, APOP, and likely
other use cases I'm not aware of.

Nobody would say replace hex. by UTF-8 in GUID/UUIDs,
therefore I fail to see the point of trying something
in this direction in Message-IDs.  Gateway operators
stating that UTF-8 in Message-IDs would be no problem
in practise, or that this would be only a minor point
when they adopt EAI, might convince me that I'm only
paranoid, but I doubt it.

And the "EAI experiment" phase did not test this plan,
there's no evidence I'm aware of that UTF-8 in Message-
IDs is harmless.  Did you really check all RFCs using
or mentioning Message-IDs for potential side-effects
of introducing UTF-8 in the "local", "unique", "left"
etc. part of a Message-ID?  E.g., does it fly with the
Message-ID concept in APOP and CRAM-MD5, or do you
intend something in the direction of "updates 2195"
to make it work?

-Frank