BitStalker is a good way to get around some of the obfuscatory methods employed by some BitTorrent trackers that are primarily used for infringing content, like the Pirate Bay. It uses some clever data-gathering techniques to avoid falling for some kinds of spoofing (like listing phony IP addresses as peers). I'm glad that some of the people at my old school, the University of Colorado, are seriously addressing the fact that current anti-piracy technology used on the Internet is inadequate.
One thing needs to be clear, though: when the paper claims that “false positives are rare with our active approach,” it's referring to false positives as to whether a particular IP address is part of a BitTorrent swarm. That is, if a particular file is known to be infringing, this method is better at identifying the IP addresses that are downloading it. A big worry of mine, though, is “false positives” that identify content as infringing when, in fact, it is not.
Imagine a work created by an author, that uses portions of a second author's work under fair use. If the first author authorizes the distribution of her work via BitTorrent, none of the people transferring that content are infringing copyright. The “authorization” of the second author is not needed, and even if the second author objects to the first author's use, there's no infringement. The statute is clear on that: a fair use is not an okay kind of infringement that we've decided to allow. It is “not an infringement of copyright.” If there's a disagreement, a judge decides–no technology should come down, by default, against free expression by making fair use difficult to actually exercise. This isn't some fantasy scenario: Universal recently lost a suit where it tried to claim that a fair use was infringing. I don't want big content companies to get through technology what they can't get through the courts.
Similarly, the “false negative” issue isn't just about getting a false negative in determining whether an IP address is part of a particular swarm. But in its discussion of false negatives, the paper doesn't address the possibility of the encryption of content, which makes determining the legality or illegality of the content impossible to those without the key. Not only does creating an incentive for encryption increase the problem of infringement, by driving it further underground, it creates the incentive for content or ISPs to use the “nuclear option” of blocking entire protocols, using whitelists, blocking or degrading encrypted traffic, and so on: “Solutions” that cause false positive problems far beyond the existing ones that BitStalker is meant to address.
(via DSL Reports)