Originally Posted by Tallest Skil
We get a lot of Polish spam, and we're starting to get spammers who hide their crap inside comments that are very nearly on-topic.
To the former, can an IP address or two be blocked to stop this nonsense?
To the latter, would modifying the sign-up process help at all?
The spam we get varies a lot so IP addresses won't block it sufficiently. When the spam got to a point where it was clear nothing was there to stop it, a filter was put in place to catch repeated posts from new users (much less sexually explicit and video encoder spam gets through, credit card fraud etc.). I keep track of spam that crops up and update the ruleset when I can - it's at over 400 rules so far and it catches a decent amount to keep the spam manageable. I have to edit the rules carefully to avoid false positives - if you filter out cialis but someone writes specialist then it catches it if the regular expression isn't right and there are also weights on certain keywords that have to be set to avoid catching non-spammers that use those words infrequently.
I believe recaptcha was put into the signup process but I'm not sure. Some spam will just have to be dealt with manually as some of the spam is in fact people not bots who sign up just to spam and it's often from China but it's fairly easy to spot when you've seen how spammers write for a few years (obviously I don't want to say which posting techniques are hard to pick up so spammers know how to spam
). The stuff that gets through pretty much gets flushed daily and often quicker.
There are other services that can be used to block spam such as akismet but it does the same as the filter we have except with our own filter, we can correct the false positives ourselves and adjust the ruleset to be specific to the spam we get. You can tell from moderators' post counts that we use the forum regularly so we are aware of the spam and I know it's frustrating having to see spam and wish you could fix it there and then but the system in place works well enough to keep the forum clean. I check the forum by recent threads so they show up pretty clearly in my normal browsing session and I check the spam reports if there's any still left.
Ideally the filter will be refined to the point where only the most obscure and harmless spam gets through but it will take time.