"Jesse D. Guardiani" <[EMAIL PROTECTED]> writes:

> How could it be implemented in the MTA though?

Ignorant of exactly how Postfix accomplishes this, I'd imagine it
would go something like this:

You receive a spam message from [EMAIL PROTECTED]'.  Before accepting
that message, your Postfix queries hotmail.com's SMTP server to see if
`bogus' is an invalid address (i.e, if mail to [EMAIL PROTECTED] would
bounce).  If so, it rejects the message, otherwise it allows the
message to be delivered to you.

In essence, only messages which can be replied to will be allowed
through, which makes a lot of sense to me.  This would significantly
cut down on the amount of spam TMDA would have to process, and also
the number of returned confirmation requests.

Does this make sense?

> I still think it would be desirable to attempt to determine whether
> a bounce is an auto-responder or a bounce at the TMDA level.

I disagree, since as I explained, this is only reliably done at the
MTA level.  You'll inevitably lose legitimate mail if you attempt
otherwise.  Not only that, but it's more efficiently done at the MTA
level as well since the verification process would be much less
expensive there.
_________________________________________________
tmda-workers mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-workers

Reply via email to