Gaming Google with Sprocket X

Posted by Simon Wharton

I saw very interesting and effective demonstration of gaming Google using a multi layered implementation of Sprocket X. Basically a Sprocket takes data from multiple sources such as Wikipedia, chews it up and spits something out that is actually meaningful and readable. It’s the readability that is the key and differentiates from other Spammy techniques. But wait, it gets really clever. So Sprocket A, if you see what I mean, pulls data in on a topic and publishes to a blog. Done well, the content iscompelling. And then Sprocket B, which is independently deployed and so has differing seed values, takes the published output from Sproket A and makes a meaningful response.. Effectively yuo get a dynamic and interesting conversation.  It gets even more interesting when humans get drawn into this spoof conversation and start responding. Effectively they start contributing and generating link love and Digg/ credibility. People start laundering the spam. I wonder whether Google is on the ball with this one.