Look into SIPs for better plagiarism detection

From Cody Gray:

I still feel like there's a way to do it with Statistically Improbable Phrase matching, which is exactly what I do manually. But maybe it's naïve to think that's straightforward to implement. For this application, it isn't necessary to match the whole post, or even score the whole post. Just flag it if you find a SIP in the post that occurs elsewhere. If you find several, the score could be increased.