Wednesday, February 3, 2010 - The Progress & Freedom Foundation Blog

Copyrights, Copycense, and Nonsense

A revealing dispute has erupted between Ben Sheffner of Copyrights & Campaigns and the not-so-competent Editors of the website Copycense, which humbly describes itself as "the online journal of code and content."

To shorten a longer story, Professor Edward Felten recently disclosed a summary of the results of a forthcoming "Sahi-Felten study" of files available to users of a "trackerless" BitTorrent-based file-sharing program. According to Professor Felten's summary, statistical analysis showed that 99% of the files available were infringing. Ben Sheffner then authored a blog post that described this summary as "[v]aluable information to keep in mind while debating net neutrality rules and IPS's right to manage their networks and fight piracy."

But Mr. Sheffner's observation outraged the allegedly pious data-prudes at Copycense. In an unsigned "Editorial" entitled Science vs. Advocacy, the crack team at Copycense thus sanctimoniously denounced Mr. Sheffner for daring to suggest that such imperfect "summary" data should ever affect important debates about network neutrality: Calling his post "reflexive" and "impetuous" they denounced his conclusion: "drawing such correlations at this point--with respect to the summary, the resulting paper, (which has not yet been vetted, reviewed, or published), or Felten's perceived or actual personal or professional biases--is premature and careless."

I will not summarize the droning Copycense account of a few of the many, many things that can inarguably go wrong during statistical analyses of sociological phenomena. Indeed, that would be pointless because Copycense itself actually concluded its sanctimonious sermon by agreeing with the substance of the conclusions that it had just denounced as "premature and careless":

"We can say with a strong level of confidence, however, that the way the current statutes are written, it would have been shocking if anything significantly less than 100% of the files on BitTorrent were technical infringements of copyright law."
How thoughtful of Copycense to admit that it knows better than to take seriously its own prudish fretting about theoretical defects that could arise from yet-to-be-reported nuances of the Sahi-Felten study. Even though Copycense currently lacks all the data that it claims to need in order to fully and completely assess all of the implications of this forthcoming study, Copycense still admits that even persons as erudite as its own Editors can, nonetheless, already "say with a strong level of confidence [that]... it would have been shocking if anything significantly less than 100% of the files on BitTorrent were technical infringements of copyright law."

If the world of legal/policy analysis had an analog to the Darwin Awards, Science vs. Advocacy would be a worthy nominee. It is the rhetorical analog of an "I'm-With-Stupid" T-shirt imprinted with an arrow that points straight up. That is how seriously this "journal of code and content" should be taken.

Nevertheless, the collective rhetoricide of Copycense still begs a question: Why did even the data-prudes at Copycense feel compelled to admit that they do not really need to scrutinize the details of a published Sahi-Felten study before they can non-prematurely and non-carelessly "say with a strong level of confidence [that]... it would have been shocking if anything significantly less than 100% of the files on BitTorrent were technical infringements of copyright law"? Well, I suspect that either or both of two factors can explain why.

First, perhaps Copycense secretly knew that the Sahi-Felten study is hardly the first statistical analysis of how piracy-adapted file-sharing programs and networks are actually used. Perhaps they also knew that its findings do not differ significantly from those of past analyses.

For example, in A & M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1013 (9th Cir. 2001), the district court found, and the appellate court affirmed, a statistical analysis that showed that 87% of the files available on the Napster network were copyrighted. Similarly, in In re Aimster Copyright Litigation, 334 F.3d 643, 650 (7th Cir. 2003), the court found no evidence of non-infringing use of the defendant's file-sharing program.

In MGM Studios v. Grokster, 454 F. Supp. 2d 966, 985 (C.D. Cal. 2006), the Court found that the distributors of the Gnutella-based file-sharing program Morpheus intended to induce the users of their program to infringe copyrights, based, in part, upon a statistical sampling study showing that while over 87% of the files available on Gnutella were infringing or highly likely to be infringing "[a]lmost 97% of the files actually requested for downloading were infringing or highly likely to be infringing."

More recently, Illinois State University's "Digital Citizen Project" conducted disclosed monitoring of campus use of a broad range of file-sharing programs and protocols. See Alexandre M. Mateus & Jon M. Pena, Dimensions of P2P and Digital Piracy in a College Campus, 1, 21, 29 (TPRC 2008). The resulting conclusions included the following:

Finally, in Columbia Pictures, Ind., Inc. v. Fung, CV 06-5578 SVW (JCx) (C.D. Cal. Dec. 21, 2009) (slip op.), the Court found that the operator of very widely used BitTorrent tracker sites including Isohunt intended to induce copyright infringement, again, based in part upon a statistical sampling study showing that "approximately 95% of downloads occurring through Defendants' sites are downloads of copyright-infringing content."

Consequently, even the summary of the Sahi-Felten study is valuable, in part, because its findings accord well with those of prior analyses. That is certainly one reason why Copycense or others could already "say with a strong level of confidence [that]... it would have been shocking if anything significantly less than 100% of the files on BitTorrent were technical infringements of copyright law."

Second, perhaps Copycense also secretly knew that the needs of lawful and unlawful commerce differ so dramatically that once a theoretically "neutral" program, network or website is widely used for unlawful purposes, that almost certainly means that it will rarely appeal to those acting lawfully.

This is particularly true in the case of what I would call "piracy-adapted" file-sharing programs, networks, and tracker-sites. As used here, the term "piracy-adapted" does not imply (until a court so concludes) that these programs, networks, and tracker-sites were intended for infringing use, but only that they happen to be fairly well suited to that purpose. In particular, such programs, networks, and tracker-sites tend to be best suited for distributing files already made popular through some other means. For obvious reasons, and with some rare exceptions, (e.g., Linux distributions), such files will inevitably tend to be infringing.

Indeed, just as the statistical data cited above would suggest, piracy-adapted file-sharing programs have actually degenerated over the past decade. Today, they tend to be less-well-suited for non-infringing uses than they were a decade ago. For example, back when Napster was the most popular file-sharing program, its centralized search-index enabled Napster users to find even rare files shared by just one Napster user. By contrast, here is a developer of the Gnutella-based file-sharing program LimeWire explaining why programs like LimeWire are actually less useful to today's garage bands than Napster was a decade ago: "Here's modern p2p's dirty little secret: It's actually horrible at rare stuff." He also notes that--albeit for different reasons--the same conclusion follows as to BitTorrent-based programs: "BitTorrent is horrible at rare stuff! As soon as a file becomes rare, it looses [sic] seeders and dies."

This may also explain why even the most sanctimonious Copycense Editors do not really need to exhaustively review every jot and tiddle of the forthcoming Sahi-Felten study in order to conclude "with a strong level of confidence [that]... it would have been shocking if anything significantly less than 100% of the files on BitTorrent were technical infringements of copyright law."

Finally, let me clarify why I find the ineptly affected data-prudery of Copycense offensive. We will never have perfectly "scientific" data about the prevalence or effects of any illegal activity--including online copyright piracy. Nevertheless, we should still want debates as important as those about network neutrality to focus upon (at least) the best available data about how piracy-adapted file-sharing programs, websites, and networks are actually used in practice. Sadly, to date, that has not been how network-neutrality law and policy have been made--as any competent Copycense Editor would know.

Consequently, only dishonest hypocrites could piously wail that some collective fall-from-policymaking-grace occurred when Ben Sheffner's blog post noted that Professor Felten's forthcoming statistical study of BitTorrent files seemed relevant and valuable to ongoing debates about net neutrality. Even if imperfect and tentative, that summary could still improve the now-dismal quality of such debates. Consequently, Copycense has tried to obscure a simple truth: It may be unfortunate that policymaking based on imperfect or tentative statistical data is often the best that we can do in an imperfect world--but it sure beats policymaking based upon mere ignorance and anecdote.

So Ben Sheffner was right--even the summary of the Sadi-Felten study can be valuable to net-neutrality debates even though it is inarguably less valuable and less reliable than a final, published study will be. Perhaps the FCC should halt its ongoing "net-neutrality" rulemaking until it can be informed by scientific studies that fully assess how the file-sharing programs that now generate most Internet traffic are actually being used. But if it does not, then it is idiocy to suggest that the FCC either can or should plow forward obliviously--ignoring all tentative or existing usage data just because it is less perfect today than it will be tomorrow.

After all, it is not as if the FCC has required prior debates about network neutrality to be informed by scientific statistical sampling to ensure that claims about how network-clogging file-sharing programs are used can be generalized from particular cases to entire population of persons using Gnutella-based or BitTorrent-based file-sharing programs. It certainly has not.

To the contrary, major debates about network neutrality--like the bizarre FCC proceeding that generated the so-called "Comcast Network Management Order"--have been driven by purely anecdotal "evidence" supplied by groups like Public Knowledge and Free Press--piracy advocates that, in Grokster, ended up defending commercial copyright piracy even more zealously than the criminal-investigation-fearing corporate pirates that they chose to defend. According to these groups, file-sharing programs are used for wonderful purposes. Gnutella? Why, Gnutella-based file-sharing programs are mostly used to "share" recordings of public-domain barbershop-quartet recordings.... BitTorrent? Why, BitTorrent-based file-sharing programs are mostly used to "share" copies of The Bible....

Even Copycense knows that such contrived stunts are wholly unrepresentative of how such programs are actually used. But here is the joke: such absurd tactics actually worked--briefly. Indeed, during the last Administration, a few contrived anecdotes actually duped three gullible FCC Commissioners into making the second-most idiotic claim yet made about file-sharing programs and networks: These Commissioners declared that although such programs and protocols had once served mostly "unsavory or even unlawful purposes," "BitTorrent and other peer-to-peer technologies, such as Gnutella, have entered the mainstream." See Comcast Network Management Order, at ΒΆ 4.

That claim was dead-wrong when made, and unsupported by any data then before the Commission. Nor has the passage of time made it less absurd--a point made quite clearly by Fung, the Digital-Citizen study, and, now, the Sahi-Felten summary.

Worse yet, the three Commissioners who made this claim were not just engaging in baseless, dead-wrong speculation. No, they were engaging in baseless, dead-wrong speculation that can mislead parents and endanger children. The Comcast Network Management Order suggested that parents need not worry if they discover that their teenager has installed a Gnutella-based file-sharing program like LimeWire or FrostWire on the family computer. To the contrary, while out-of-it old people might dimly recall that such programs were once used mainly for "unsavory or even unlawful purposes," federal experts at the FCC knew that by 2008, "peer-to-peer technologies, such as Gnutella, have entered the mainstream."

But that was and is horribly wrong. Worse yet, it is wrong in ways that can end promising careers, ruin families, destroy businesses, increase health-care costs and fraud, compromise national security, endanger the wife and children of President Obama--or even end the lives of children in the worst possible way. Gnutella-based file-sharing programs never "entered the mainstream"--available data suggests that they have always been used overwhelmingly for "unsavory or even unlawful purposes" like piracy, identity theft, and distributing child pornography. Consequently, the Sahi-Felten summary can already contribute something valuable to the "net neutrality" debate--just by reminding federal officials how badly they can mislead and endanger the public by assuming that contrived antics concocted by piracy-loving "public interest" groups actually represent "mainstream" file-sharing.

Copycense thus spouted nonsense. Fortunately, its exercise in self-contradiction has ensured that its attempt to conceal advocacy of continued ignorance and anecdote beneath a pious veneer of respect for the integrity of "science" was too incompetently executed to do any real harm.

So I am not criticizing the crack team at Copycense because they want policymakers to know that they intend to go over the Sahi-Felten study with a fine-toothed comb in order to expose any of "[Professor] Felten's perceived or actual personal or professional biases...." Fine: Have at it.

Rather, I am criticizing them because--given that even they admit that we can already "say with a strong level of confidence" that it would be "shocking" if the summarized conclusions of the Sahi-Felten study failed to be correct--then nothing but grotesque hubris seems to explain why Copycense could claim that responsible policymakers must ignore the Sahi-Felten summary and study and keep basing decisions about net-neutrality laws on ignorance and anecdote until Copycense bestows upon humanity whatever contributions to "science" can be made by persons who cannot keep their own story straight during one blog post.

posted by Thomas Sydnor @ 1:56 PM | Copyright , Cyber-Security , E-commerce , IP , Internet , Mass Media , Neutrality , Think Tanks