Another spambot/spammer is attacking BLF. They can be pretty annoying. So that people later will know what I am talking about, this post is all over our forum at the time of this writing:
EDIT: I just broke all of the links. Every part in bold was a link to their website.
Is this how CPF began the journey to the darkside? Idiots spamming and CPF trying to come up with ways to stop it. I hope we are able to kill the spam without losing the community feeling we have here.
Hmm. Next step would be to create a technical solution to prevent spam from being accepted.
For instance, generate a hash of each message over some minimum length (so :P messages don't trigger). Then just keep these hash signatures in memory for a few hours (to keep the collision checking quick) and dump any new messages that collide.
You could defeat this by changing a single letter of the comment, yes, but it would defeat pure cut/paste attacks.
Good idea! There are implementations of algorithms that are designed to find near-matches (such as simhash and probably lots more, considering that there's a lot of interest in academia). Simhash, in particular, is actually pretty fast, relatively speaking of course.
Also, while we're talking site issues, I thought I'd bring up this thread. NeoGeo discovered that a blf.cc.cz link in an old post of mine now points to a domain squatter site. NoScript and a custom filter list actually blocked the redirection attempt in FF but when I opened the site in IE (in a separate sandbox), it redirected to a spam site. I suspect this affects all old links that point to blf.cc.cz (ie. the old site). Is there any way to re-write old links in the database so that they point to the new site? Getting the old subdomain back would obviously be easier, but I suspect the spammers won't be too happy to give it up and cc.cz is notorious for ignoring abuse cases.