In draft-ietf-marid-core-03.txt change the second paragraph of Section 3 to
read:
vvvvvvvvvvvvvv
This question will usually be asked by an SMTP server as part of deciding
whether to accept an incoming mail message. However this question could also be
asked later, after the server has accepted the message, or by a different party.
An MUA, for example, could use the result of this question to determine how to
file or present a message.
^^^^^^^^^^^^^^
and insert immediately below that paragraph a new paragraph...
vvvvvvvvvvvvvvv
Where the question is asked by an SMTP MTA server after that MTA has accepted
the message for onward transmission, the server shall itself generate and send
non-delivery messages ( in accordance with RFC3464) derived from the rejection
codes detailed in section 5 below.
^^^^^^^^^^^^^^^^^^^
and add a normative reference to RFC3464 in section 10.1
=======================
Rationale
The existing draft describes tests as being undertaken during the RFC 2811
message-acceptance process.
There may be MTAs in which it is desired to accept the message when offered, and
then undertake the tests at a later time, perhaps after reputation sources have
been consulted and other tests applied.
The current draft leaves MTA behaviour in post-acceptance tests undefined.
It seems right-and-proper to require that the overall handling of a message by
the MTA-chain be consistent, regardless of whether pre- or post-acceptance
testing is chosen.
These proposed changes to the draft are intended to ensure that, so long as the
message is still within the SMTP transportation process, coherent behaviour is
exhibited, regardless of when the Sender-ID test is actually undertaken.
Chris Haynes