There is a lot of talk about the spam problem. Tim Bray suggests a system of paid email that depends on a passworded SMTP server that forwards email anonymously in order to eliminate the spam problem.
There are three reasons this doesn't need to happen:
1. Most of the spam problem resides with sys admins who either allow spam to be delivered through their servers or that leave their SMTP relays open. The former should be dealt with through blacklisting -- they already are. The latter need to be taught through some kind of negative reinforcement (how about... blacklisting?) to close down open relays.
2. Virtually any server should be able to screen for spamming where it starts. Today, we leave it to the recipient to screen out spam and a few ISPs claim they block the stuff. However, email servers, everything from Microsoft to the SMTP MTAs that ship in Linux and UNIX (including the Mac OS), should include a default setting that prevents large numbers of email with the same or similar subject lines from being sent -- this way, the admin would have to "choose" to be or enable a spammer.
3. Dozens of routing protocols can be used to screen spam near the source and the technology exists to screen every packet for the fingerprints of spam using Bayesian and adaptive filters. If the people who pay for the bandwidth being used by spam want to stop the wasting of their network capacity, the cost of adding this capability can be recouped quickly.
Push the responsibility back onto the sources of spam, not the end-user who generally doesn't spam one iota. Frankly, I don't mind the spam I get, because, like Tim Bray says, only a few dozen get through a day and I can delete them, adding the addresses and subjects to my spam filter. It's a small price to pay relative to the potential cost of the 12,000 to 15,000 email messages I send in a year (don't laugh, that's just an average of 33 messsages a day), which under a one-cent plan would cost me an additional $120 a year.
Thanks to for the links.
UPDATE: I neglected to note this yesterday -- Don Park's approach, based on using spam filters on SMTP servers with fixed IP addresses, is a good start, especially because it creates a feedback mechanism that improves server-side filtering.
Posted by Mitch Ratcliffe at October 13, 2003 09:15 AM | TrackBack