ばか！バカ！　馬鹿ー月！

Apparently bots have been registering accounts without us noticing back four months past and then coming back assaulting the wiki hard. I banned nine accounts total.

Having said that, is there a way to introduce image text recognition into any anonymous edits *or* registration?

we have to petition to the big boss, velocity!

send him an email and he will get right to it, it'll also help if you send him a link to a mod he can simple download and install on the wiki if you know where it is.

I'm not convinced they are bots. If they're people with nothing better to do, a CAPTCHA isn't going to do squat.

One of the dirty hacks I did on a forum I was running... spammers would often register and put a spam link in their profile. You could set up your profile before being approved or e-mail verified, so the reg didn't even have to complete to get their link into the system.

So what I did was tweak the code... near the "Website" field I put a note: "Leave this blank when registering, you can change it later."

If the field was filled in, then I knew it was a spam bot and I silently stopped the registration process (blank page). This was 100% effective in stopping the spammers, since they could no longer use automates bots to spam links and manual attempts would be caught at the verification stage.

I bring this up because, perhaps, we could figure out a way to analyze the contents of the data being posted and block it if it meets certain criteria, such as links to known spam sites. Even if they obfuscate the URLs, that wold make them invalid and the search engines wouldn't count the cross-linking (plus users wouldn't be able to click them directly). That effectively defeats the purpose of the spam.

Just a thought, I'm not sure how involved that would be.
=Smidge=

your right. i've noticed the bots posting the links in the website field.

But from velocity post i thought he was talking about the wiki, not the forum?