To try and help stem the tide of spam comments here, I recently upgraded my version of Movable Type and am going to instal the MT-Blacklist app, which tries to automatically block spam.

But for a while there, I was fighting the spam simply by manually blocking the IP addresses of any spambot that posted here. The spambots tend to post in enormous packs: One will find my a blog, discover that it's open to post to, then leave about 200 postings in the different items. So if you're the blog publisher and you see the first few postings go up by the 'bot, you can quickly ban its IP address and shut it down.

But that means you have to be doing a very quick, on-the-fly Turing Test: You have to be able to quickly look at a post and decide whether it looks like it was posted by a human or not.

I recently failed that test! Alfred Cloutier, a regular poster on the Collision Detection boards, was noticed that a spambot from an online gambling site had posted some spam. Alfred decided to intentionally post his own spam-like comment -- he used his real name, but his comment was like one of the typically vague-but-friendly spambot sort ("nice site thx"). I saw it and automatically banned his IP, because I was moving so quickly I didn't notice the name of the poster. Alfred emailed me to point out my error, and I unblocked his IP address, and also felt like an idiot.

It is, of course, yet another example of the difficulty of living in a world filled with so much pseudointelligence that you can't recognize real intelligence.