When I read this post:
PN7 was exposed mainly on linguistic proof. FOMA worked hard to dispute it. I'm quite sure that this was an attempt to discredit such proof in case this type of link was ever established between his accounts.
It gave me a crazy idea: would you be able to add
Statistical Methods for Natural Language Processing? It would be very cool to get the result for any 2 usernames selected!
I know, I know, it's very easy to come up with things that take a lot of work to implement, and a large amount of processing power to crunch the numbers. But I had to ask!