"Meng" == Meng Weng Wong <mengwong(_at_)dumbo(_dot_)pobox(_dot_)com>
writes:
Meng> The Caller-ID algorithm for selecting the Purported
Meng> Responsible Address/Domain was derived empirically.
I really don't think it was. It follows logically from the
definitions of the From, Sender and Resent-* headers in (2)822.
I would have proposed pretty much the same algorithm for SPF, in place
of just looking at the From, but then scope=header went away, so there
was no need.
If Microsoft has tested this on BIGNUM real world messages, to
ascertain that we don't have so many systems failing to follow (2)822
as to make it useless, then that is of course a good thing.
I do know of at least one MUA that doesn't follow the RFCs, though,
and this will result in failures with Caller ID and with the New SPF.
Eudora has always supported resending messages but (at least in older
versions) it didn't add Resent-* headers. Such messages will of
course fail chacks that are based on the PRD.
-roy