Spammers begone!
Posted on March 23, 2004 @ 2:52 pm
The comment spam here at Apropos of Something finally got to be more than I could handle, so I installed Jay Allen’s MT-Blacklist earlier this afternoon. Peddle your generic Viagra and hentai elsewhere, spammers!
I had heard that MT-Blacklist was easy to install, but I had no idea that it was going to be that easy. It seems to be working fine, but let me know if you encounter anything that’s broken.
Posted by Jess | Filed Under Blickity-Blog |
4 comments so far...
Leave a Comment
If you would like to make a comment, please fill out the form below.
I am not so sure about MT Blacklist… any sort of blacklist is gonna be prone to human error. and overzealous entries in the blacklist can cause problems down the road…
I am really loving MT-Bayesian, though. It’s getting pretty good at recognising spam… well, the few I get anymore, as I have changed most of my “Comments:” or “Your Comments:” references.
You know, I’m normally a bit skeptical of blacklists myself, but MT-Blacklist had a couple of factors going for it. First, its blacklist is based solely on URLs. Therefore, the only way that a spam comment is going to get blocked is if it links to a known e-commerce spammer (and the spammers have to have the link, obviously, for comment spam to work). In that sense, there’s less room for error than if it were based on IPs. If a spammer isn’t on the blacklist, the comment can admittedly still slip through. If that happens, though, it’s easy enough to add the URL to my local blacklist. Better still, MT-Blacklist includes a de-spamming feature that will search through all the comments on your blog to find and delete any comments recently added to the blacklist. Hence, it adds quite a bit of functionality to Movable Type comment management in general.
The other factor that made me want to give MT-Blacklist a chance is that hundreds (if not thousands) of MT users already have it running. If a problem pops up in the blacklist, there’s a decent chance that someone will notice and report it. Since MT-Blacklist logs all its blocked comments, it’s easy to see when there’s something amiss.
How much spam did it take to train MT-Bayesian?
I’d say about ten to fifteen or so… mine will still sometimes err on the side of false-positives, putting legit comments in as spam, but it’s an easy interface to retrain.
I haven’t had spams show up on my page, even if they get posted, in a about a month now.
According to my logs, MT-Blacklist is 3 for 3 so far. Rawk.