Many observers of the fight against spam bemoan "collateral damage", the blocking of "valid" emails, but much of this criticism is misplaced. Though content-based blocking (alá SpamAssassin) is likely to block valid emails with no upside, the same cannot be said for the blacklists based on pinpointing insecure or spam-friendly servers. Blacklists like ORDB or MAPS RBL+ block based on mailserver IP address, and in these cases, "collateral damage" is a feature.

The only way to pressure people who run insecure servers to clean up their acts is to make it painful for them to be bad internet citizens. Perhaps they don't care that spammers are using their computers to relay trash to everybody else, but if people they care about stop accepting their mail, perhaps they'll be moved to act.

Most sites that have open relay have no idea, and they only find out when ORDB blacklists them. Then their mail is refused, and those who look at their mail logs can see who's gotten bounced. I wrote a small ordbscan tool for just this purpose, and I have a customer who's gotten quite a few business partners to secure their machines. This would not have happened without the blacklists.

If somebody complains about content-based blocking eating their mail, this is fair, but whining about IP-based blocking mostly reflects self-inflicted damage.