Tony,
TF> The reason Sender-ID falls back to checking MAIL FROM is because SUBMITTER
TF> may be omitted if it is the same as MAIL FROM. If a Sender-ID
TF> implementation is checking the MAIL FROM in the absence of SUBMITTER,
TF> should it use the v=spf1 record or the v=marid1 record?
One of the more interesting errors that we all have made in designing
protocols is the pseudo-efficiency hack of using defaults.
With unfortunate frequency, I keep seeing this "assumption" about
values come back to bite us. From/Sender is merely one example.
(Another is "default" character set for MIME, but that's another
diatribe.)
Might be nice for a new protocol to reflect something from the learning
curve, and avoid this mistake. No Submitter field should mean no
submitter value in rfc2821 headers, rather than 'look at mailfrom'.
d/
--
Dave Crocker <mailto:dcrocker(_at_)brandenburg(_dot_)com>
Brandenburg InternetWorking <http://www.brandenburg.com>
Sunnyvale, CA USA <tel:+1.408.246.8253>, <fax:+1.866.358.5301>