Mark Andrews wrote:
It was more "the only way to prevent the A lookups is
to remove the fallback".
If we ever head down this path there would need to be
years of advance notice.
My € 0,02: There is no direct way to get rid of the old
A-fallback, 2821bis can say "please" in a note, maybe an
additional BCP could go as far as SHOULD, and that is it.
Indirectly the great master plan AFAIK is that IPv4 will
die, and without IPv4 the A-fallback also dies. So what
we now have to decide is if we want an AAAA-fallback for
IPv6, or if we have a chance to kill this AAAA-fallback.
Any "mail opt-out" alternatives like "v=spf1 -all" or
the nullmx draft are odd if the great master plan is to
have names and IPv6 addresses for "everything", notably
billions of things never intending to send or receive
mail.
A "mail opt-in" for IPv6, where only names with an MX
are interested in mail, would be more straight forward.
The "opt-out" alternatives can play a role while IPv4
still exists. But they would be an ugly kludge for the
brave new IPv6 world.
Frank