Flashes of Panic

Everything I do at work to filter spam from our incoming email is "free software" ("Free as in speech," they say, "not free as in beer.") There's a bit in Wired News today which underlines that decision. Two new filter developers (using various refinements on mathematical recognition of spam fingerprints) are claiming Ivory-soap levels of accuracy in identifying spam, which is a good thing on it's own. But what I find remarkable is the reaction of the "professional" spam-hunters like Brightmail:

"People can make any kind of claim at any time," said Francois Lavaste, vice president of marketing at Brightmail. "You can make claims today, but what matters is how they hold up down the road."

Even if independent tests prove CRM114 and Dspam to be more effective, Lavaste cautions potential users to consider whether they need the training and support that vendors of commercial solutions can provide.

"ISPs might find it attractive and acceptable, but is it an ISP-class solution?" said Lavaste. "That remains to be seen."

Here's the problem: Brightmail depends, for its very existence, on a continuing flow of spam to justify the expense of hiring them. What is their incentive to completely shut out spam? It's like the Soviets; once they were in power, where was their incentive to achieve "true Communism" where "the state would wither away?" Brightmail doesn't want to wither away, and they don't want open-source spam filters cutting in to their business. Whereas the open-source spam-filtering guys would rather get back to doing something really interesting, instead of shoveling manure. Who would you trust to get the job done properly?