At 10:23 am +0100 2004/05/07, Andy Fletcher wrote: >Giuliano,
>
>To clarify, my question was about sender callouts. I assume from the
>discussion this means the changes pointed out do not affect sender
>callouts? If so, what is the likelihood of this changing, by either way
>I suggested or alternatives - there are a lot of broken mailservers
>out there (by design or mistake) which reject a null MAIL TO but which
>you'd rather accept mail from than bounce unconditionally just because
>of this reason.
>
>I think there was some historical discussion on this, where someone
>said that they don't see the primary use of sender callouts as being
>to prevent bounces being undeliverable, but in todays world the major
>use it combatting spam/UCE/general clutter. I couldn't agree more - if
>you're doing recipient verification and rejecting during the
>transaction then the only bounces you're likely to get are from full
>mailboxes and the like, which should be rather minimal, since most
>rejection is done during the transaction.
>
>Andy
Andy, I agree almost completely. My server would generate a bounce
only for failed forwarding (I believe mailbox full would generate an
immediate failure, wouldn't it?). I disagree only on the need of
controlling sender for sender callouts, I think this should always be
the null sender, but I think that a more fine grained response is
needed, as I pointed out before.
