Re: Mailing lists - assumptions

> On 19/04/2014 19:00, Ned Freed wrote:
> >
> >> The way alum.mit.edu handles email forwarding for life is they provide
> >> an authenticated submission service for MIT alumni[1].  This handles SPF
> >> and DKIM on the outgoing side, and on the forwarding side, it looks to
> >> the e-mail provider that that an address such as "[email protected]"
> >> is a mailing list.
> > This isn't the issue; the issue is what the MAIL FROM address gets set to
> > when a message is forwarded.
> >
> > You want any downstream delivery errors to be reported to the original sender,
> > so ideally you'd leave the MAIL FROM unchanged. But now an SPF check on the
> > MAIL FROM will fail because alum.mit.edu's IP address isn't going to be
> > allowed list.

> Do you want delivery errors to be reported to the original sender? Are
> you sure?

Of course you do. We're talking about a pure autoforwarder here, not
a mailing list. The only mailboxes in the picture are the ones belonging
to the sender and the recipient. Who else does it make sense to report
the problem to?

Mixed situations, like those created by having a sieve-enabled mailbox, are
another matter. There's no hard-and-fast rule for MAIL FROM manipulation that
produces optimum results in all cases.

> If I send a message to '[email protected]', do I really want lots of
> delivery errors from random people I've never heard of? Wouldn't it make
> more sense for those delivery errors to be reported to the list
> administrator?

Now you're talking about a mailing list, where overriding the MAIL FROM
is always the right answer. Indeed, overriding the MAIL FROM in some
senses the defining characteristics of a mailing list.

In any case, this is different funcitonality than an autoforwarder provides,
with different semantics.

> I know we have this problem with our MTA where people send mail to
> '[email protected]', and that user has a forward to a gmail account. It
> confuses the original sender if they receive a delivery failure report
> from Google, because they never sent a message to Google at all.

That's why we have the NOTARY extension and the ORCPT field. Its purpose
is to preserve the original recipient address so it can be presented to
the user when there's a nondelivery.

I also note that the ORCPT field is reset on MAIL FROM changes. Quite a lot of
careful consideration has gone into this design, you know.

> So, we
> normally set things up so the MAIL FROM is changed to go to
> [email protected]. That way, if a forward is set incorrectly,
> they can sort it out, and the company's contacts don't get confused.

In the case of a pure autoforwarder operated, say, on behalf of a former
alumnus, who is this magical administrative person who is both willing and
appropriate to see nondelivery notifications containing all or part of the
original message supposed to be?

An anecdote. Back when I was doing the postmaster thing, I was once braced by
an extremely angry user who had received a message from a postmaster at an
intermediate site who had taken it upon themselves to perform the role you're
talking about here. In this instance they forwarded a piece of mail that had
gone awry, but also saw fit to attach their own commentary on the content.
(Amusingly, the postmaster who did this is still around, and is now a very
vocal advocate for Internet privacy. I smile every time I hear the name.)

Trust me, this is not a situation you want to enable.

				Ned

_______________________________________________
ietf-822 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ietf-822