ITH success in any Web site comes some spam, which needs to be combatted effectively. Herein I will deal with social bookmarking Web sites in particular. Spam is not always automated. There is brute-force spam that is scripted; but there’s also self-promotion that strikes in the form of mass submission. And people have begun trying to game Netscape, whose front page bears an admirable PageRank 9. This keeps the Anchors and Chief Editor on their toes.

The problem has become somewhat universal across this new wave of sites comprising contributer-driven content. Digg has had submission parasites, yet human moderation, as well as spam report widgets (community-driven), have it eradicated early on. I usually report any suspicious submission as spam, at least as soon I spot a distinct and objectionable pattern. When the same person always posts to the same domain, for instance, that’s a red flag. Sometimes you can align the username with the domain’s affiliation, but sometimes consistency in the URL is enough. And ‘sock puppets’ (same person with multiple identities that boost a bogus sense of consent) are another-yet-closely-related matter altogether.

When enough stories get intercepted, links the to the domain are banned by principle (for a month if not permanently), or particular Web addresses blocked for good. This sends the appropriate message: abuse, then get your domain blacklisted. This may be better than banning the users who could otherwise change their ways and contribute differently; in a positive way, that is. In fact, some people just haven’t grasped the concepts of social bookmarking, so they fail to see the wrongdoing.

When banning users, there is a need for caution. A pissed off innocent user is far worse than spam that successfully percolates because people talk. They have blogs, so a good rant with proof can get heavy exposure very quickly. And it affects reputation. Look at what has happened in Digg more recently.

An afterthought: One possible workaround to ‘sock puppets’ would be to demand that each newly-subscribed user supplies a unique E-mail address, as well as logs in with an IP address that wasn’t yet used in registration on that same day. This can’t stop instantiation of puppets or protect against proxies and dynamic IP’s. However, it definitely slows down the abuse and reduces incentive to game the system.

This entry was posted on Sunday, September 17th, 2006 at 12:40 pm and is filed under Blogs, CMS, Social, Spam. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.