Andrew Newton writes:
I completely disagree. Mailing lists are nothing more than
special types of forwarders.
I don't think there is any consensus in the broader IETF community on
the role of a mailing list. It is not an MTA but it is also not an
MUA. I've seen these arguments before and this has never been
resolved.
Which is why I really don't think of this in terms of MTA's,
mailing lists, etc, etc -- trying to tease out precise
definitions is so many angels on a pinhead. Instead, I think
of this in terms of agents in the transmission path which
might want to take some responsibility for some identities
for whatever reason they might have. The home domain's
motivation is generally to take responsibility for the From
address. A mailing list MTA might want to take
responsibility for the Sender (or other) addresses. It may
have further motivation because it mangles the message in
such a way that previous signatures are trashed [*]. Thinking of
it this way, we neatly avoid both the need for definitions
of what functionality a thing is (and the inherent
schizophrenia that brings), as well as this procrustean
desire for a single party to take *complete* responsibility
for the message when in fact it a shared responsibility of
those participants in the forwarding path. Best of all, the
receiver decides for itself how it wants to weigh the
relative worths of the asserted identities; the receiver's
local policy should have final say-so given the information
gathered in the forwarding path, not some edict on high from
IETF.
Mike
[*] again, we should make the default stance is that
manglers SHOULD play nice with MASS, not the other
way around.