A brilliant idea? Or is it fraught with errors...

[Kareeser]

Random thought du jour...

Ideally, spam bots want to maximize their efficiency, and would prune inactive emails from their servers.

Ideally.

So what if I spoofed the address typically used by qmail (like UNDELIVERABLE@kareeser.com), and manually bounced the message back with the message typically used by qmail (or whatever)?

Theoretically, the machine would be programmed to eliminate the address from the database... worst case scenario would be that the machine figures this out and sells my address to more people...

Basically, my only spam comes in packages of ~13 a day, all with similar subjects, so I'm fairly sure they're from the same set of machines...

Of course, if the remote machine is using a spoofed address, or a randomly generated email address, then I can't exactly do anything (the whole double-bounce bit), but... would it potentially work?

[Lespaul20]

I think your last sentence answers your question. I'm no expert on spam bots but I would think that if they are using other computers besides their own that they couldn't care less about being efficient. I really doubt they try to keep a legitimate e-mail database since most spams use e-mail address to send from that are used and then thrown away so the odds of them actually getting a undeliverable is probably pretty slim. I would just stick to using spam filters.

[mairving]

Some spam filters do something similar. They will bounce back messages as undeliverable if it is from a known spammer.

In your case I wouldn't do it because most likely it would be more trouble than it is worth. I don't think that spammers are really into maximizing their efficiency much anymore. It's quantity over quality nowadays.

[faulkner132]

You would be putting an unnecessary load on your mail server by generating and sending (probably to a dead address anyway) all the NDR messages.

Just have good server side or client side spam filtering to delete the messages.