Barry Leiba wrote:
The only concern I have here is that because "user agent" has a
specific connotation, there could be confusion about what happens to
it when a user uses more than one UA. Suppose I use Gmail's web
client, Mulberry, Apple Mail, and Thunderbird, all at different times,
and all sending mail as <barryleiba(_at_)computer(_dot_)org> through Gmail's
email infrastructure. (In fact, that *is* the case.) Should the UAID
that Gmail sticks in the signature always be the same, because it's
just me, using these client programs interchangably?
I think this mostly confuses implementation with architecture.
Simple test: Are you using/getting different email addresses when you use
these
different modes of access? Probably not.
That said, note that John Levine's description of his signing model nicely
provides an example of having different UAID values, depending on the author's
means of access. In this case, he's decided that there is in fact some real
utility in making exactly the distinction you raise. So it's plausible, but no
it's not necessary.
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html