Tuesday, May 1, 2007

You CAN Copyright a Number. But Not As Such.

Back in February, a hacker named Arnezami made public (on the Doom9 forums) the HD-DVD Processing Key. And the Industry-with-an-evil-capital-I is very upset over that.

Basically, the HD-DVD Processing Key is a special number, necessary (but not sufficient) for the decryption and watching of HD-DVD's on your home Linux box. Which seems fairly innocuous - after all, once I buy a disk, by rights I should be able to watch it on any screen in my house. I use the phrase "by rights" on purpose; the question is clearly covered by the first sale doctrine.

But note, however, that once it's been decrypted, it can not only be watched but pulled apart, copied, redistributed, uploaded, downloaded, sideloaded, whatever. So you can understand why the publication of the information might upset players in an industry with a business model built on, you know, not allowing that to happen so much. Even if the information in question is just a very big number.

Now, per Slashdot, which points to Rudd-O, which used to point to Strange Action at a Distance, but that blog's been squashed... the Evil Industry is attempting to stomp out the information. Apparently, (quoting Rudd-O) "the movie industry is threatening Spooky Action at a Distance for publishing that number, specifically with copyright infringement." According to the quote that Rudd-O pulled off of Spooky Action, the threats are premised on the DMCA.

Slashdot, Rudd-O, Spooky Action, and possibly others are all speaking in terms of the DMCA copyright provisions; There is already some hue and cry to the effect that you can't copyright a number! (Even a number with letters in it, as when converted to hex.)

Now, first and foremost, it's very likely that the threat was not based on the copyright provisions in the DMCA but rather on the anti-circumvention provisions. And that should be the end of this discussion.

But it's worth taking a step back for a minute, and addressing the question of copyrighting a number. Because you actually CAN copyright a number. Just not if you're trying to protect it qua number.

That is, if I figure out some awesome new method for figuring out the next huge Mersenne prime, and then write that up in an article, then under copyright law: I can NOT protect the method I describe; I can protect the article in its particular expression, unless the method is only describable in a limited number of ways, in which case I can't; and I can NOT protect the number itself. Anyone else can come along and point out that the Nth Mersenne prime is 12457...whatever, and I can't stop them. And if they want to represent that number in Hex, it becomes a string of numbers and letters, and I still can't stop them.

HOWEVER. I can protect a number when it's not a number. I remember years ago someone wrote a poem that was made up of fictional personalized license plates. Which would look just like a long hex number, except it's not a number qua number. Or the line from Shakespeare: "2B, R not 2B" On its own, probably too short for copyright protection, but not by dint of its being essentially a string of letters.

I would even claim that if someone wrote an abstract poem, consisting of nothing but a string of numbers and letters that would sound "right" when read aloud in sequence, that would be protected under copyright.

On a more technical level, every document saved on your computer is, in a very real sense, nothing more than a number. It's stored as a series of bits and bytes which the computer generally translates as words or images; but there's nothing that obligates computers to do so. Just for fun, take a small image (it has to be small so as not to overwhelm your computer) and open it in a text editor (Notepad or the not-Windows equivalent). It's a bunch of gobbledygook, but it's something.

(Just a brief nostalgia moment: I remember the heady days of Usenet, when big files were broken up and posted as long, apparently meaningless text files full of hex codes that you'd concatenate and then rename to have your image or music or document or whatever.)

The point is, you can protect a number, in some circumstances. In the instant case, the number is probably being used qua number - it's the answer to a math problem and so is probably not protected under copyright. But then, I reiterate, copyright probably isn't the issue here; the threats are probably being made under the DMCA anti-circumvention provisions.

As importantly, though, what you also can't do to a nubmer is to make it go away by wishing it so. Back in the Summer of 1995, when I worked for the EFF, I helped out (in a one-L summer intern kind of way) with the Bernstein crypto export case.

In that case, I recall, the algorithm in question could, with some rejiggering, be represented as a string of numbers a few lines long. A few times - most often in the spirit of sincere legal analysis and only sometimes in the spirit of stick-it-to-the-man humor - the question would come up as to what the law would (and should) be if, instead of seeking to "export" the source code and an academic paper about the algorithm in question someone had just put it on a t-shirt and flown out of the country wearing it.

Of course, those conversations took place back in the day when if you wanted something a little more professional than using a Sharpie on your Hanes, there were some fixed costs involved. Now we have cafepress. Plus ca change, plus c'est le meme chose. Some young clever has even registered the number as a domain name, and has a nascent blog there.

So, as a copyright question - not protected. And as a practical matter - not protectable any more. I didn't go into the trade secret question, but that's also not going to work to keep the number out of the public's grubby mitts.; Arnezami apparently figured out the code by looking at disks, not by gaining improper access to guarded information.

My first instinct under the anticircumvention provisions is that release and propagation of the number is problematic, but as I write these words, I'm less and less convinced. I'll have to think about it some more... maybe even post a follow-up. It'll come down to the question of whether or not the Processing Key is an "access control" as used by the statute.

And looking at the broader question of copyrighting numbers...

You can copyright a number. Just not this one.

--Ben D. Manevitz

Update: Lots of good coverage out there. Only one link to me, but them's the wages of sin. Or in this case the wages of toiling in obscurity. In any event, I found particularly heartening the coverage at EFF Deep Links, that makes clear that it is, in fact, an anti-circumvention claim, and not a copyright one.

Also - tagging is a very cool and excellent innovation in the blog world. For those of you who don't check the tags on posts you read, it might be worthwhile to check the tags on this one at least. I got it from Slashdot, and I'm guessing I'm not the only one.

Labels: , ,