Come get your p*ker and v14gr4 -- spam is being nuked
Jeremy just released the new version of his Drupal spam module. Yes, it's true that Drupal sites suffer from spam as well -- it's just always had a comment moderation system, so you can easily place comments from anonymous users into an approval queue.
This works great at low volumes. For high traffic sites that get lots of high traffic spam...it doesn't work at all. Legitimate comments get drowned out by a flood of spam -- e.g. I have 36,000+ comments sitting in my approval queue. But, since I've been an advocate of open comments, I didn't want to close things down and force people to register.
So, tougher measures are called for, and at 10:30PM PST tonight, the first piece of p*ker comment spam bit the dust. And I didn't even need captchas. Thanks, Jeremy!
Similar
Comments
Results
Looking forward to hearing the results of the nuking, and to know what kind of stats the module generates. I installed captchas on my site (more to test that it was possible than anything), and so far zero spam, but very early in the game. Registered users are exempt. I'm a weblog spam skeptic, influenced mostly by Mark Pilgrim's article on the subject, and I'm especially skeptical of technological solutions that can be gamed. That said, open source attempts are interesting because the algorithm is by definition public, so the rules are out in the open.
It's pretty funny that this was marked as spam :P
Maybe for mentioning "spam" too many times? There is pretty extensive logging, but not stats. It was actually the bayesian filter that caught your comment -- everything else has just been hitting the custom filter for various spammy words like those mentioned in the post title.
So...never mind complicated bayesian stuff, at this point I just need comments that are obviously spam to get nuked automatically. I also had 4 other legitimate comments come through with no problems, where I would some times miss them in the flood of spam.