LawMeme LawMeme Yale Law School  
LawMeme
Search LawMeme [ Advanced Search ]
 
 
 
 
Can XOR Eliminate Copyright?
Posted by James Grimmelmann on Wednesday, June 09 @ 12:51:29 EDT
Contributed by Anonymous
Copyright
Anonymous writes "What happens when you take two copyrighted, binary files and "munge" them together, bit-by-bit, using the logical XOR function? You end up with a third file that contains, by itself, none of the information present in either of the original files. Is the resulting file copyrighted? If so, by whom?

The munged result is certainly not a direct copy of either original file, nor is it a derivative work (since it contains no part of either file in recognizable form). Of course, the munged result can be combined with either original file to construct the other file, which makes the legal distribution of the munged file dubious. Is this an uncopyrighted file that is illegal to distribute according to copyright law?

We have entered the uncharted (and murky) backwaters of the digital copyright debate. Monolith is an open source tool that can be used to explore these issues.

Note:

(JG): It's an interesting idea, but unfortunately, I don't think it's a hard case for copyright. The Copyright Act defines "copy" in a way that I think includes Monoith XORs:

''Copies'' are material objects, other than phonorecords, in which a work is fixed by any method now known or later developed, and from which the work can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.
17 U.S.C. s. 101. "Material objects" doesn't cause a problem, since the material object is the hard drive on which the munged file is stored. Same goes for "fixed:" that copies in RAM are "fixed" for purposes of triggering copyright is, sadly, established law. This leaves only "from which the work can be perceived, reproduced, or otherwise communicated."

The problem here is that you can perceive the work from a Monolith munge: just XOR it again and then pipe the result through the appropriate app (e.g. a sound player for audio, less for text, etc.) There's not really a major difference between this case and the case in which the file is sitting on your hard drive in its normal format and you pipe it through the appropriate app: both involve "perceiving" the work "with the aid of a machine or device."

One might object that the presence of the second file means that the munge isn't a copy of either of them. This has intuitive appeal on the definition of "copy" as we use it in everyday English, but I don't think it works as a matter of copyright law. First, it's my understanding that as a matter of copyright law, "copy" and "derivative work" aren't exclusive: the same new work can be both a copy of and a derivative work of the old work.

Second, this kind of transformation is more or less indistinguishable from encryption: it involves the use of a "key" to fiddle with the bits of a file in a reversible way. (What makes Monolith "different" is that it's symmetric between key and cleartext.) But encryption, as a general matter, isn't problematic for copyright: it's just another format for storing a work, like the choice between MP3, OGG, and WAV, which are just different forms of fixation and require different techniques to play. I don't know of anyone claiming that MP3 files or password-protected ZIP files aren't "copies:" if that argument would fly, I'd have expected one of the file-sharing services to have advanced it in court.

Finally, if you could produce an uncopyrighted work by XORing a copyrighted work with some other text, you could eviscerate copyright law. We could all just agree to use a common text -- Down and Out in the Magic Kingdom, say -- as the other half of our munge, and we could ship around and trade any file we wanted. Putting on my legal realist hat, I can't imagine a judge standing for that kind of result; therefore, she would interpret copyright law in such a way as to avoid reaching it. Therefore, XORs are copies.

Note: I saw this submission in the queue this morning and was intriguied enough to sit down and bang out this reply on the spot. I haven't done any legal research on the matter and am not willing to swear I'm right. I've mangled my copyright law before on this site under similar off-the-cuff circumstances, so I should definitely not be taken as the voice of authority. Still, I don't think it's possible to hack copyright law this easily."

 
Related Links
· More about Copyright
· News by James Grimmelmann


Most read story about Copyright:
Top Ten New Copyright Crimes

Options

 Printer Friendly Page  Printer Friendly Page

 Send to a Friend  Send to a Friend

Threshold
  
The comments are owned by the poster. We aren't responsible for their content.

Re: Can XOR Eliminate Copyright? (Score: 0)
by Anonymous on Wednesday, June 09 @ 13:38:39 EDT

This is a trivial case of "transformation", which (at least in the UK) is codified in statutory copyright law as an infringement of the rights subsisting in a work.

The original poster wrongly believes that "nor is it a derivative work (since it contains no part of either file in recognizable form)", as "recognisable form" is not an issue: what is an issue is that the resulting work was _derived_ from the original work: a direct chain of copying can be shown.





[ Reply to This ]


Re: Can XOR Eliminate Copyright? (Score: 0)
by Anonymous on Wednesday, June 09 @ 15:06:27 EDT
I agree with your take on this XOR method of obfuscation. It doesn't seem any different than ZIP or even ROT13. It seems fairly obvious that such a file would just be considered a copy as it is merely junk on it's own.

What would happen though, if the "munged" file were not merely junk? For example, what if mp3 files were encoded with english words, resulting in a new and unique text? It wouldn't be the most space efficient method of encoding, but would it be able to stand as it's own creation?



[ Reply to This ]


Re: Can XOR Eliminate Copyright? (Score: 0)
by Anonymous on Wednesday, June 09 @ 15:49:01 EDT
I think this is just a case of our side making the same mistake that the other side has raised to a science: trying to find a technical solution to a non-technical problem. Mathematics doesn't solve copying, and it won't solve copyright, because those are not mathematical problems. At best it can provide a support to the human-realm solutions.

Now, being that the math side of these things is actually my area of expertise, I'd take issue with one of Anonymous's claims:

You end up with a third file that contains, by itself, none of the information present in either of the original files.

Depending on the nature of the original files, it may be more accurate to say that it contains ALL of the information of BOTH the original files. The entropy (that is, the mathematical uncertainty) of a file is often much less than the number of bits in its representation. For instance, English-language text has an entropy rate of a little over one bit per character. That means that if you have a text with one letter unknown, on average you can guess the missing letter and have about 50% probability of being right. One bit per character is a lot less than the eight bits per character of standard ASCII text files.

So if you were to take two ASCII text files and XOR them together, there'd still be a whole lot of redundant information in the result. The resulting file might look odd on the screen, but would only have about two bits of entropy per byte. With appropriate software it would be possible to recover both texts with a pretty good degree of accuracy - not perfect, but probably good enough. Every so often someone proposeses almost exactly this system as an encryption scheme in sci.crypt; they're always flamed soundly for its insecurity, and then either learn from the experience or spend some time blustering about it.

Now, I realise that Monolith isn't being proposed as an encryption scheme as such, but as a tool for exploring "the boundaries of copyright". I don't think it'll be very useful for that, though, until a lot more thought has been put into both the encryption and copyright sides of what it's trying to do. I don't find the points made on the Sourceforge page covincing at all. At first glance, it seems pretty clear to me that the output of Monolith is a derivative work from both its input files. Whether you can or can't recover the original from the derivative, or whether the derivative contains any "useful" information, are not interesting questions from a copyright point of view because those were never the defining attributes of derivative works.

Imagine for a moment that I take a photo of a copyrighted painting, but defocus the camera so that all is visible in the photo is a smear. Doesn't matter. It's still a photo of a copyrighted work of art, and still a derivative. If I wanted to be sole owner of copyright on my smear, I should have made it with something public domain as the input; if I claim that that would make a difference, then I'd admitting that my work is a derivative.

Matthew Skala
What I'm thinking about
CS PhD student, co-author of The Breaking of Cyber Patrol 4


[ Reply to This ]


Re: Can XOR Eliminate Copyright? (Score: 0)
by Anonymous on Wednesday, June 09 @ 17:37:18 EDT
Consider: Files A and B are XOR'ed to produce file C. Does the owner of file A then also own file C? If so, then the owner of file A owns all other files in the world because there is always some file B with which it could be XOR'ed to produce some file C, and vice versa (truncating or repeating the process as needed to match lengths, etc). Logically and mathematically, it has to be so.

Judges, however, are not bound by logic and are free to rule on the basis of political correctness to obtain whatever result is desired. I believe this is sometimes called results-oriented jurisprudence. So it is probably a moot point to argue the logic involved because such a case would surely be decided based on the identities of the parties involved on a case by case basis.

For example: If I, as an common individual, were to claim ownership of a particular music file being distributed a record company by claiming it could be produced by XOR'ing it with another file, one that I own, most judges would probably rule against me. Reverse the roles and the judge would probably still rule against me. Politics requires it, irrespective of logic.



[ Reply to This ]


Re: Can XOR Eliminate Copyright? (Score: 1)
by bodo on Friday, June 11 @ 08:44:17 EDT
(User Info | Send a Message) http://www.warsystems.hu
Hello All! If You accept a comment from a newbie, Monolith does not do anything different from what has been done with DeCSS code [www-2.cs.cmu.edu]. That did not stand before judge Kaplan. One interesting question for me is what happens when the result of a technical transformation or transcoding process is something in the public domain, like a prime [primes.utm.edu]. I guess there are similar methods to transfer any digital file into anything else, i mean there must be a mathematical function that is able to transform windows C source into war_and_peace.txt both the function and the novel per definitionem in the public domain...


[ Reply to This ]


Brightnet chunks are not encryption (Score: 0)
by Anonymous on Tuesday, June 22 @ 13:51:10 EDT
Finally, if you could produce an uncopyrighted work by XORing a copyrighted work with some other text, you could eviscerate copyright law. We could all just agree to use a common text -- Down and Out in the Magic Kingdom, say -- as the other half of our munge, and we could ship around and trade any file we wanted. Putting on my legal realist hat, I can't imagine a judge standing for that kind of result; therefore, she would interpret copyright law in such a way as to avoid reaching it. Therefore, XORs are copies.
That's why you shouldn't use one single common XOR. In Monolith terms, you should take your mono files and make them into a basis file. A xor B produces a public domain work like The Constitution of the United States A xor C produces a copyrighted work like "Cryptonomicon". B xor D produces a public domain work like "The Tempest" B xor E produces a copyrighted work like "Fight Club" and so on. Every given chunk of data is just a big number, that can be xor'd with another into something public domain or something copyrighted. You'd have to search someone's computer to prove a user had produced something unauthorized with it. That's how brightnets are supposed to work, according to The Big Hack [www.thebighack.org].


[ Reply to This ]


Finally figured this out - my (Score: 0)
by Anonymous on Tuesday, July 20 @ 09:38:32 EDT
While the lawsuit claims that Schwarzenegger's repeal of the car tax was illegal, others have argued the opposite position. Senator Tom McClintock has argued that former Governor Gray Davis's tripling of the car tax was illegal. Order Utilities [order-utilities.fateback.com] Order Internet [order-internet.fateback.com] Order Business [order-business.fateback.com] Order Games [order-games.fateback.com] Order Development [order-development.fateback.com] Order Graphics [order-graphics.fateback.com] Order Screensavers [order-screensavers.fateback.com] Order Homehobby [order-homehobby.fateback.com] Order Audio [order-audio.fateback.com] Order Miccelaneous [order-miccelaneous.fateback.com] While the lawsuit claims that Schwarzenegger's repeal of the car tax was illegal, others have argued the opposite position. Senator Tom McClintock has argued that former Governor Gray Davis's tripling of the car tax was illegal.


[ Reply to This ]


freaking BS, I enjoyed the IP (Score: 0)
by Anonymous on Tuesday, August 03 @ 08:09:35 EDT
Before raising the car tax, the state also has to be in a position where it is unable to borrow money, said McClintock, and neither was the case when Davis increased the tax. Snore Stopper [www.makehelp.com] Apnea [www.makehelp.com] How To Stop Snoring [www.makehelp.com] Snoring Children [www.makehelp.com] Obstructive Sleep Apnea [www.makehelp.com] Sleep Apnea Products [www.makehelp.com] Tips To Stop Snoring [www.makehelp.com] Snoring Products [www.makehelp.com] Snore [www.makehelp.com] Cures For Snoring [www.makehelp.com] Before raising the car tax, the state also has to be in a position where it is unable to borrow money, said McClintock, and neither was the case when Davis increased the tax.


[ Reply to This ]


Leges humanae nascuntur, vivunt, moriuntur
Human laws are born, live, and die

Contributors retain copyright interests in all stories, comments and submissions.
The PHP-Nuke engine on which LawMeme runs is copyright by PHP-Nuke, and is freely available under the GNU GPL.
Everything else is copyright copyright 2002-04 by the Information Society Project.

This material may be distributed only subject to the terms and conditions
set forth in the Open Publication License, v1.0 or later.
The latest version is currently available at http://www.opencontent.org/openpub/.

You can syndicate our news with backend.php



Page Generation: 0.292 Seconds