On 12/5/22 22:35, George Michaelson wrote:
If a message had an EXPIRES-BY header, semantically past it's "due"
date, and an intermediate store-and-forward SMTP MTA decided to look
in its queue, would it be "wrong" for it to return to sender rather
than deliver? If the header could be shown to be origin specified, and
not arbitrarily added?
(note that EXPIRES-BY is not the proposal under consideration)
In general I would say that it's poor design to specify anything in a
message header that intermediate MTAs are allowed or expected to act
on. Partly this is because the header is at the wrong layer. One
example of a problem that can result is that messages can be resent from
one recipient to another.
Expires: should IMO be understood and interpreted as part of the message
content. Sure, a recipient can delete a message based on Expires, just
as a recipient can manually delete any message that the recipient sees,
or write a sieve or other script to do deletion based on any pattern in
the message. That message is the recipient's to do with as they
please. But if a message is delivered, maybe archived, and maybe resent
somewhere else, the original Expires header field is part of the
content, and context, of the original message. The recipient to whom
the message is resent might again decide to delete it, but no
intermediate system should make that decision. Even an "expired"
message is still useful for some purpose, just presumably not the
original, primary purpose of the message.
(and sending the archived message as an attachment is not the same
thing; it has a different meaning)
Keith
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
