Copyright

[tyme]

For details of current copyright terms, see http://www.copyright.cornell.edu/resources/publicdomain.cfm


Suppose I have a copyrighted recording, image, ebook, or piece of software.  Let's label it M.  I create a random string, R, of bits, with the same length.  I xor the two together to obtain M' (m-prime).

Suppose the copyright owner forbids redistribution.

Now suppose entity X is distributing R, and entity Y is distributing M'.  A bunch of people download R and M'.  Who is violating copyright, and why?

Along with your answer, please state which copyright term length you support.  :)

[Nick1911]

Who is violating copyright, and why?

Personally, I think solely the end user, presuming they recreate the original work from the two derived attributes.

Legally, probably everyone involved.

Along with your answer, please state which copyright term length you support.

Lifetime of author.  I believe the author is entitled to the fruits of his or her labor.  However, I think this entitlement should end when the author dies, and the work becomes public domain.

A good thread.  I look forward to seeing your answer, tyme.

[MechAg94]

I didn't know we had to do algebra on this site.   

[AZRedhawk44]

So you want to encrypt digital media, then share the media and a decryption key from two different sources.  Neither on its own is the source material.

Except, M' is based on the source material and could not have been created without it.

The goods you mention (images, books, software, music, movies, etc) are funny things.  Some might be spewed out very quickly with little effort on the part of the creator, while others may be a lifelong labor.

Establishing a timeline on which one can derive sole profit for such a good is difficult as a result of the difference in time it takes to create it.

Then, you add in the wrinkle of corporations owning copyrights rather than individuals.  Corporations don't die, and even if they disincorporate, the copyrights are tangible assets that are sold off.

My initial response is copyrights should last no longer than the lifetime of the copyright holder.  In the case of corporations holding copyrights, they should last no more than 20 years.

This view was put in place for me by JRR Tolkien's forward in a copy of Lord of The Rings that was printed back in the 70's.  Evidently another publishing house (Ace books) found a loophole in international copyright and started printing his books.  I forget the exact wording of the forward, but the part that hit me was along the lines of:

...Those who believe in courtesy to living (at least) authors will buy this edition, and no other.

In my mind, the creator of M' (who also either created R or used R as a random string of bits for an encryption key) is the copyright violator.  And anyone who utilizes M' and R to re-create M.

[makattak]

Personally, I think solely the end user, presuming they recreate the original work from the two derived attributes.

Legally, probably everyone involved.

Lifetime of author.  I believe the author is entitled to the fruits of his or her labor.  However, I think this entitlement should end when the author dies, and the work becomes public domain.

A good thread.  I look forward to seeing your answer, tyme.

I think his analysis is correct.

However, I think about 20 years should be the length of time for copyright. We need incentive to create, but I don't think the entire lifetime of the author is necessary.

I also think our current laws are FAR out of whack. I could live with simply the lifetime of the author as well. However, I think his heir(s) should have some claim if the majority of the value of his work has not been realized. Thus, I think 20 years may be a better result as I would think the majority of the value should have been extracted.

[red headed stranger]

I think y would be the most at fault.  


I think 20-30 years after the authors death is reasonable.  This would make it possible for an author's estate to help support any dependents that they may have at the time of their death.  

[zahc]

Since copyright is nothing but protectionism, I don't think there should be copyright at all.

[Nick1911]

Except, M' is based on the source material and could not have been created without it.

My problem with this as a litmus test, is, say:

This is some text created by Nick1911!

MD5("This is some text created by Nick1911!") = 48414b60e679af0e8fbd5682074019d3

Does this mean I own the copyright to 48414b60e679af0e8fbd5682074019d3?  After all, it's extremely improbable that anyone could create that without the source material.

[AZRedhawk44]

Well, the number 17 isn't in there, so you're safe.

 

If you just posted a web page with the MD5 results of MD5("This is some text created by Nick1911!") and didn't offer an explanation that the contents of the page (or a txt file at a link) was supposed to be "This is some text created by Nick1911!", then it would be harmless and no-fault.

If, however, you posted a page with the MD5("This is some text created by Nick1911!") and the decryption key, with the intent of clandestine (or overt) distribution, then the copyright holder of "This is some text created by Nick1911!" would have claim in my mind.

If it is ADVERTISED and a decryption key is MADE PUBLICLY AVAILABLE then it is redistribution.

If it is not advertised and no decryption key is publicly available then:
1.  No one knows what it is and won't sue;
2.  No one can reconstruct it anyways.

If I took the Return of the Jedi DVD and ripped it to ISO, then MD5 encrypted it and then published a link to a file named RotJ_md5-48414b60e679af0e8fbd5682074019d3.iso, it's pretty obvious that the string at the end is an MD5 decryption key to make the file use-able.  That's a violation in my mind.

[tyme]

The problem is that in terms of information theory, AFAICT, R and M' are the same.  They are both random.  Only in combination do they have any relation to M.

The fact that M' exists makes R just as "connected" to M as M' is connected to M.

If, however, you posted a page with the MD5("This is some text created by Nick1911!") and the decryption key, with the intent of clandestine (or overt) distribution,

Couldn't the legal argument be made that by publishing anything on a website, you are presumed to have intent to distribute?  I agree that whether someone else knows that R or M' is makes all the difference in the world, but I can't imagine any court in the real world delving in the nuances of entropy enough to make that distinction.  That's akin to trying to judge the legality of something based on the result.  I don't think that works.

[AZRedhawk44]

Except what you're doing is really just adding another layer on top of an existing work.

Open up an ISO or mp3 in a text editor.

That doesn't "look" like anything use-able, either.  Just a bunch of random characters.

Except, when you run it with a CD burning suite and put it in a DVD player, or play the mp3 with WMP, then the appropriate codec interpreter comes into play.

If I have a movie that has been converted to a Matroshka codec rather than the original DVD codec, it's still that original movie.  Just compressed/decompressed differently.

You're altering the steps necessary to utilize the information, but the information is still the same.

[tyme]

Neither R nor M' contain enough information to reproduce the song, no matter what encoder is used.  Alone, they contain no information about the song.  If you have one piece, the entire other piece is necessary to reproduce the song.  Even though they are based on each other, if you only have one piece, you need either the complement or M itself in order to "have" the song.

I don't think I can agree that the process of creating M' merely "alters" M.

[AZRedhawk44]

Neither R nor M' contain enough information to reproduce the song, no matter what encoder is used.  Alone, they contain no information about the song.  If you have one piece, the entire other piece is necessary to reproduce the song.  Even though they are based on each other, if you only have one piece, you need either the complement or M itself in order to "have" the song.

I don't think I can agree that the process of creating M' merely "alters" M.

Let's say you've never heard of the matroshka codec.

I take a movie you created.  I rip it from DVD to matroshka.  I put it on the desktop of your computer.

You have no idea what it is.  Your computer doesn't have the matroshka codec.

It's still that same movie, though.  Just because you don't have the codec (it's the parallel to R in your example) doesn't mean it isn't the movie.  You just haven't discovered HOW it is the movie.

[AZRedhawk44]

I don't think I can agree that the process of creating M' merely "alters" M.

Yes you can.  You even admitted it with your initial logic.  You named it M'.  Because you know it is derived from M, and is procedurally dependent on M.

Your initial suggestion is no different than:

I have a file.  It contains:

Code: [Select]

That Redhawk fella is awesome!

Taking that file, and find every-other bit in it (the odd ones) and moving it to a second file, while substituting a 0 for the stolen bit.  You end up with:

file1:

Code: [Select]

0h0t0R0d0a0k0f0l0a0i0 0w0s0m0!

file2:

Code: [Select]

T0a0 0e0h0w0 0e0l0 0s0a0e0o0e0

If you have a means to reconstitute "That Redhawk fella is awesome!" then it's no different than winzip, tar, cab, ISO or any other means of file compression, encryption of obfuscation.  It's still the same source data when you re-assemble it.

The point is, you will strip these obfuscations off and re-assemble the file.

[Nick1911]

Except it isn't just obfuscations unless you own both pieces.

Say I give you this:

1YZ2XI)XSV_ESFTH]C9GEC@EsEYR*XKUJAL]

What's it mean?  Is it a copyright violation?  How is this distinguishable from random noise?

Without some other part, it's meaningless.

[Tallpine]

That Redhawk fella is awesome!

'S e fear-iongantach a tha'ann Seabhag Dearg!

[AZRedhawk44]

Hey, why'd you edit my post, Tyme?

[roo_ster]

Personally, I think solely the end user, presuming they recreate the original work from the two derived attributes.

Legally, probably everyone involved.

As far as liability, I think Nick191 hit it.  Morally, the end user, legally, all of 'em.

I voted "16-30 years."  More precisely, I would run with, "Author's lifetime or 20 years, whichever comes first."

[makattak]

As far as liability, I think Nick191 hit it.  Morally, the end user, legally, all of 'em.

I voted "16-30 years."  More precisely, I would run with, "Author's lifetime or 20 years, whichever comes first."

I'd use "Author's lifetime or 20 years, whichever comes last" OR simply 20 years (I'd prefer just 20 years).

He has the right to pass his intellectualy property onto his heirs. That's part of property rights.

[tyme]

Sorry, I goofed up, hit edit instead of quote.  I edited it back.

I don't think I can agree that the process of creating M' merely "alters" M.

Yes you can.  You even admitted it with your initial logic.  You named it M'.  Because you know it is derived from M, and is procedurally dependent on M.  

No I can't.  Suppose the random generator function generated M' instead of R.  Then R becomes the "derived" work and M' is free and clear.

Just because M' was derived from M in the initial example doesn't mean it has any different status than R in terms of information content / entropy.

[BrokenPaw]

MD5("This is some text created by Nick1911!") = 48414b60e679af0e8fbd5682074019d3

Well, the number 17 isn't in there, so you're safe.

Sure it is.  Twice.  And the second time, it's backward, thus conclusively proving obfuscatory and therefore nefarious intent.

That's constructive possession, that is.

[HankB]

Patents are good for 20 years, so I don't see why a basic copyright should last longer . . . in some cases, MUCH longer.

[Nick1911]

Sure it is.  Twice.  And the second time, it's backward, thus conclusively proving obfuscatory and therefore nefarious intent.

That's constructive possession, that is.

You made me laugh out loud. 

Yea, that's totally an obfuscated way to write 17.  How can I prove otherwise?

[280plus]

Does anybody else's head hurt? 

 

[Tallpine]

Yea, that's totally an obfuscated way to write 17.  How can I prove otherwise?

How about: seachd deug ?