So now that the spammers have figured out how to hit this fourm, is Jive considering any options to deal with it?
Capatchas are popular, but anoying to real people. A one time thing for account signup wouldnt be too bad, though- would it?
What about DNSBL? I dont know if fourm spammers are using the same netblocks as the email spammers, but it might be worth considering.
Filtering? It would be interesting to see some sort of filtering done. Even the basic word filters could catch a lot, using a score based system like SpamAssassin would be better (it would be cool if you actually put real support in for SpamAssassin).
How about a "Report this as Spam" button and let majority rule? If more than 60% of the views of a message report it as spam, delete/remove the post. Or some other metric determined with imperical testing.
Of course, using a range of technologies makes the most sense. Since the spam problem has been tackled fairly well for email (though not universally implemented) I think you could employ some of those methods here.
I like the idea of a user having a spam score (not just the message, but the user). The spam score can be calculated like the Bayesian method with a common database, so the number of posts a user has is relevant. If your first post talks a lot about the stock market, those words weigh quite heavily agaisnt your spam score. But if after a few hundred posts you mention some stock item it only marginally changes your score. And for efficiency you could stop worrying about the score after a certain number of posts and just "whitelist" the user.
How are other public fourms dealing with this problem?