Amusingly, even after reading the whole paper I was unable to determine if it was just a clever collection of puns or if actual experiments had been conducted (and then written up humorously.) "Imperative detection" certainly seems like a plausible method, for one thing.

What I'd really like to see, as a future direction of research, is a filter that can automatically differentiate between free galleries and ad splashes for gateways. ;)