\n"; echo $styleSheet; ?>
Home > Copyfight

June 09, 2004
Monolith - An Uninteresting Experiment in CopyrightEmail This EntryPrint This Entry
Posted by Ernest Miller

BoingBoing links to a new "copyright experiment" (Monolith and digital copyright). The software project, called Monolith, takes two digital files and XOR's them (what the author refers to as "munging"), creating a third file. The author calls the two input files "element" and "basis." I think many people might call them "plaintext" and "key." The output file (aka the "monolith" file) would be called the "cryptotext."

The conceit of the concept is that neither the cryptotext nor the key is copyrighted. Thus, it should be legal to distribute both. Otherwise, the author of Monolith claims, everything is copyrighted and nothing can be distributed because there is always a number such that, if XOR'd with another number, will produce a copyrighted work.

This argument is not new and it not terrible interesting. It basically postulates that any encrypted transmission of information is actually not a transmission of information at all.

LawMeme: Can XOR Eliminate Copyright?
Joe Gratz: Monolith: Cool Idea, Doesn’t Work

Toehold did the work I was too lazy to do and has a brief history of the concept of evading copyright this way It's still rockin' XOR to me.

Category: Counterpoint

Seth Finkelstein on June 9, 2004 02:50 PM writes...

What you said!

Let nobody ever argue to me that getting mentioned in the BigBlogs is a matter of being good. It's a matter of being coming up with what they want to hear, no matter how inane.

Permalink to Comment
Terry Posey on June 9, 2004 07:58 PM writes...

Perhaps I've missed the forest for the trees, but isn't the "mono" file a derivative work (under 17 U.S.C. 101), which is just a different type of infringement?

I guess the Mono author is arguing that because you can't "see" the original work in the new file, it's not, but I think it still falls under a technical reading of a derivative.

Permalink to Comment
Brad Hutchings on June 9, 2004 08:44 PM writes...

So what would you say about a system where 100 people form a little club and each excerpts exactly 1/100 of a copyrighted work. Then each could use an automated method to query for the 1/100 part and reassemble the excerpts to get the whole work.

Or how about a decryption or circumvention device that works similarly? Would that be inane enough?

Permalink to Comment
Ernest Miller on June 9, 2004 09:08 PM writes...

Terry, the encrypted file is plausibly both a derivative work and a copy. I might even say it is simply a copy and not a derivative.

There is more to a copyright analysis than your example provides but, yes, one could imagine copyright infringement using such a system.

Permalink to Comment
Terry Posey on June 9, 2004 09:27 PM writes...

Well, I agree that digitization poses copyright problems (the Mono author's underlying thesis), but I think they are more the result of ease of distribution rather than the metaphysics of copying.

I think Brad is describing the problems behind suing Bittorrent users for infringement. How much financially can (or should) I liable for if I sent 1/100th of a work to only 3 people, 10 people, or 100 people?

Permalink to Comment
Crosbie Fitch on June 10, 2004 04:58 AM writes...

Remember, even an MP3 file is not a copyrighted work. It is a means of reproducing a copyrighted work (just as a piano roll is).

Thus a file of apparent garbage and a PIN are just as much infringing as an MP3 file.

It is a separate question as to whether one infringes a work simply by disseminating the means to reproduce it (as opposed to disseminating an actual reproduction).

Now the issue of garbage and key is that it raises the question as to whether data quantity should be involved in determining infringement (or the degree thereof). If I possess a 5M MP3 file without the suffix 'mp3' and I just need to remedy this in order to reproduce a piece of music, how much different is this from having a 4 digit PIN and simply needing 6M of encrypted file?

If I downloaded 50 PIN codes would I have infringed 50 works simply by possessing part of the reproduction instructions for each?

The issue is whether infringement can occur even in part. Once we've established that then we can decide whether an infringement requires that at least say 90% of the data must be possessed.

But, then we get into the strange situation of compression. If there is an MP3 file of a song that is 90M, and also a version that is 9M, how does anyone determine whether my 9M file is 99% of something or 9%?

One must catch the user reproducing the work red handed.

Or perhaps it would be best simply to abandon copyright in the digital domain.

Permalink to Comment
Mike RoChannel on June 10, 2004 06:24 AM writes...

I think you completely miss the point.

The question is not about encrypting/hiding content.

The question is: If both original files are copyrighted (could be also Pictures) and you mix them producing a third file, that in no way looks like either one for source files, who owns the copyright of that result file?

See, as in the "Gray Album" you can mix stuff and still the copyright holders claim that the end result is theirs. But if you mix up using Monolith there is nothing original left. And still you are aber to reproduce one of the orginal files using the other ...

Permalink to Comment
Crosbie Fitch on June 10, 2004 08:56 AM writes...

If you inspect copyright in the digital domain too closely of course you'll realise it's a bunch of nonsense. However, one of the key issues is not what's obvious on inspection, but how the law perceives the matter. Unfortunately, the law is happy to be an ass with regard to copyright and arbitrate over whether five angels are permitted on a pinhead.

Encryption means transforming into an obscure form - typically reversibly. Using XOR does this.

If you create a product P of two files A and B. Then supplying P with the instruction "XOR with A" is a copyright infringement of B, because you are providing information on how to reproduce B.
Similarly, supplying P with the instruction "XOR with B" is a copyright infringement of A, because you are providing information on how to reproduce A.

If you supply P on its own and with no clue that it may produce anything else, then sure, perhaps you haven't infringed any copyright - yet.

Now, let's say you feel you have the copyright on P. If you do, then people infringe your copyright simply by disseminating the instruction "XOR A with B", i.e. providing sufficient information to reproduce P. But, by doing that, one has simply demonstrated that your work derives from two others with your contribution of using XOR to create a new work. So, you can't defend your copyright if you ever reveal how P was produced.

So, there remains the question: What if people shared files of apparent noise that would only potentially become copyright infringements if small pieces of additional information were to be subsequently provided? Thus, people share umpteen proto-DVD copies for a few weeks, and then everyone stops sharing those, and instead starts disseminating the tiny decode keys?

Alternatively, if it could be ordained that copyright infringement required sharing at least 10% of a file, everyone could share 5% fragments of every file under the sun. Whenever anyone wanted to reconstitute a particular file they'd just download the nineteen other 5% fragments from 19 different sharers (from many hundreds having those same fragments).

The point is, where does copyright infringement happen? The first bit? The hundredth? The millionth? or the whole file?

Or do we create a new law? "Conspiracy to commit copyright infringement"

The digital domain was designed for copying. Copyright in this domain is like a 'prevention of cruelty to pixies' act in the physical world.

Permalink to Comment

TrackBack URL: http://www.corante.com/cgi-bin/mt/mt-tb.cgi/3028
A new "copyright experiment" . [Read More]

Tracked on June 15, 2004 04:01 PM

Fun with copyright from MemeMachineGo!
Illegal Primes: What folks often forget is a program (any file actually) is a string of bits (binary digits)—so every... [Read More]

Tracked on August 30, 2004 05:25 PM





Remember personal info?

Email this entry to:

Your email address:

Message (optional):