Justin Mason sums up the timeline on this paper and various responses between the authors and the DSPAM developer. I'd put Cormack in the top 3 professors who taught computer science when I attended Waterloo.

Spam: or, 'SlashDot spam drama'. So, a few days ago, I forwarded a link to a paper I'd been sent -- it's a great paper, and I'm not just saying that because SpamAssassin did well -- it really tests some of the popular open-source spam filters comprehensively, and correctly. (The authors have 24 years of information retrieval research between them.)

The results have been pretty incendiary. ;) Here's a timeline with links, in case you were wondering where we are right now:

• I forward A Study of Supervised Spam Detection Applied to Eight Months of Personal Email, Gordon Cormack and Thomas Lynam, to the SpamAssassin-dev mailing list for some quiet review. ;)

• it hit Slashdot (that was quick) entitled 'Spamassassin Beats CRM-114 In Anti-Spam Shootout'. A bit over-the-top, but that's Slashdot for ya.

• The DSPAM author responds with a SlashDot post entitled Response to Gordon Cormack's Study of Spam Detection, linking to 'an appropriate response to Cormack's technical errors' -- with a few errors of its own.

• Some follow-ups: Henry Stern's, my own, and Lynam and Cormack's.