--On Thursday, 27 March, 2008 12:31 -0700 "Hallam-Baker,
Phillip" <pbaker(_at_)verisign(_dot_)com> wrote:
The key issue here is whether people who rely on AAAA are
likely to achieve their desired result. Today it does not
matter because anyone who relies on AAAA alone with no A
fallback is going to receive almost no mail.
This is true iff you believe that there are too few IPv6-enabled
mail senders to transmit such mail. It ultimately has nothing
to do with whether there are MX records present, since having an
explicit MX record that pointed to a mail exchanger host that
only had a AAAA record would leave the sender in exactly the
same situation -- deliver over IPv6 or don't deliver at all.
That parallelism moves the discussion entirely out of the scope
of whether the text in rfc2821bis is correct or not.
More important, I suspect that belief is false. Installations
who run IPv6-only generally know their correspondents. The
correspondent would presumably need to be running dual-stack, or
have a submission server or equivalent that is running
dual-stack (or IPv6 only as well), but, in many parts of the
world, those are not rare.
A statement I believe would be true is that a mail server that
only supported incoming connections over IPv6 would receive
almost no spam, at least until IPv6 becomes much more popular.
But, the last I heard, that would usually be considered an
IETF mailing list