I already made a well-working CAPTCHA patch to mlf month ago, but Alex
showed no interest!

Yes, I did! But in my opinion CAPTCHA is no good solution as it affects the usability. Therefore I'd like to implement a spam protection of which the (human) users don't take notice of. I think a central and up to date filter system like Akismet is a good approach. But at the moment I'm also testing Bad Behavior and I'd appreciate any other good idea.

Alex