I think a lot of commenters are missing an important point of the article. The whole thing was the scientific process in a nutshell: he made an observation (most blog spam is ie6), made a couple hypotheses (most ie6 traffic is bots, blocking ie6 will significantly reduce spam), tested them out, and shared the results. He was even able to make a different, related conclusion (numbers for windows and ie6 browsing are inflated) based on his work.

Overall, I found the piece quite interesting. I think you will, too, if you read it not as a piece of advice but as a nicely-performed experiment.