On Sep 19, 2007, at 2:35 PM, Martin Zobel-Helas wrote:
Hi, On Wed Sep 19, 2007 at 14:09:32 -0700, David Brodbeck wrote:That's true, although the sending server will generate a bounce back to the sender. So the mail doesn't disappear down a black hole, at least.Great! with million senders being forged every day that helps a lot.
Pretty much all the spam I see these days is sent direct-to-MX by trojaned PCs running dedicated spam-spewing software. Rejecting at SMTP time doesn't create any backscatter spam in this situation, because the spamware isn't going to bother to generate a bounce!
Accepting a message and *then* bouncing it would indeed generate backscatter spam and would be the wrong way to go about it, of course. Once the SMTP transaction is over a message should never be bounced, but it's OK to refuse to accept it at SMTP time, IMHO.
Older Exchange servers can be a major culprit in backscatter spam. I discovered a while back that Exchange 5.5 accepts *anything* at SMTP time, even invalid usernames, then creates a bounce message later. This is utterly broken. Some Linux MTAs can be configured this way, too, and misguided folks sometimes implement spam filtering this way.