A New York district court has ruled that space-saving measures used by music locker service MP3Tunes.com were within the bounds of copyright law, Ars Technica reported yesterday. The decision [PDF] is a complicated one, and Ars does a great job parsing it all out, as usual. Suffice it to say that both parties got something out of it, although MP3Tunes and its founder, Michael Robertson, were found liable for some infringement.
There’s one part that I find particularly interesting. MP3Tunes uses a method called hashing to compare files its users attempt to upload. Identical files produce identical hashes. This allows them to use a unique hash for every song and “recreate” that song for everyone who uploaded it to their locker. This process is called de-duplication because the company doesn’t need to store an actual copy of the file for each user who uploads one, and it’s a great way to save disk space and bandwidth.
Previous case law suggested that using a single “master” copy of a song to enable all users to listen to or download it may constitute a copyright infringement. While that view may withstand additional scrutiny, Judge William Pauley said that’s not how MP3Tunes serves its files.
Instead, when “a user plays or downloads a song from a locker, the system uses the hash tags associated with the uploaded song to reconstruct the exact file the user originally uploaded to his locker.” Hence, Judge Pauley appears to say, hash-based file reconstruction is not infringement because each user is served a song created from the signature of the file they originally uploaded.
Amazon and Google should be very happy about this portion of the ruling. Right now, neither company has a deal in place with record labels for their music services. Each one allows users to upload their own MP3s, and that approach insulates them (at least so far …) from copyright liability. But they don’t de-duplicate, so when you and I both upload a copy of ”The Most Beautiful Girl (In the Room)” by Flight of the Conchords, Amazon and Google store both copies on their servers. Now, imagine music collections, all saved in their own folder, hundreds or thousands of times. That’s a lot of disk space that this ruling could save for both companies and others like them.
If this ruling were the end of the matter, both Amazon and Google could reclaim some of the massive amounts of disk space they recently devoted to their music services. But it’s bound to go to appeal, and there’s no telling how it will all pan out. Consequently, I don’t expect either company to risk changing their approach to this problem until a more bullet-proof decision comes out from a higher court (or the highest court).
This case, however, illustrates the truly square-peg-round-hole nature of applying copyright law to modern technology, even in the wake of updates like the Digital Millenium Copyright Act [PDF]. I think Nilay Patel, formerly of Engadget and now at This is my next/The Verge, said it best during a recent podcast rant. I’m paraphrasing, but his point was that copyright law regulates copies instead of content. That problem is the primary source of frequent, counterproductive and complex copyright litigation. And that litigation will continue until the dissonance between what is regulated by copyright law and what should be regulated by copyright law is actually corrected.