Slashdot Mirror
Author Drops Copyright Case Against Scribd Filter
natehoy writes "Apparently, monitoring for copyright violations is not in itself a copyright violation, lawyers for Elaine Scott decided. As a result, they have dropped the lawsuit against Scribd, who was being simultaneously sued for allowing copies of Scott's work to be published, and retaining an unlicensed copy of the work in their filtering software to try and prevent future copyright violations."
This is getting really frickin ridiculous.
They are being sued for not blocking copyrighted data, and then sued for holding a copy in their filter so that they can block further copies? WTF?
What do you even say to that kind of idiot?
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Of course, I knew that already. My hard-drive contents, including the "diff" program are one big copyright violation monitoring tool.
If Pandora's box is destined to be opened, *I* want to be the one to open it.
1 - forbid site from avoiding copyright infringements.
2 - sue when copies of copyrighted works appear on site.
3 - PROFIT!
It's a new strategy for pirates.
"I'm not pirating software! I'm watching out for copyright infringement, and I need a copy in of the pirated product in order to do just that!"
Let us suppose we are an image hosting site, that has in the past been used to host child porn (Think Flickr or ImageHost). By the same logic, it would then be appropriate for us to maintain copies of child porn in order to filter new uploads against it. In my opinion, the only organizations that should be allowed to retain copies of c.p. are those government organizations actively involved in policing it -- regardless of motive. So when our company gets v&, do you think they'll accept our filtering excuse? Notwithstanding that the IP laws are screwed up, It was still an illegal copy, and I feel the author's case has merit.
What about monitoring monitoring for copyright violations? Is meta-monitoring a copyright violation?
Except that your #1 is not the facts of the case.
The site was using an unauthorized copy of the work to check for other unauthorized copies.
Stealing a car to look for stolen cars doesn't make you a cop.
I buy the DMCA safe-harbor defense, that seems iron-clad to me, but they were still knowingly making copies of her works (even if they didn't make them available) and profiting off the information. I would argue that they had a pretty good chance of winning on that claim.
Still this is really a non-story, "person with lawyers decides not to sue", sounds like something the Onion would run.
Playing a devil's advocate a little bit here but is there no way to avoid copyright infringements other than storing the entire copy of a copyrighted work in your system?
Negative moral value of force outweighs the positive value of good intentions.
I swear, judge! I downloaded that copy of Backyard Vixens 7 because I was trying to make sure nobody infringed on their copyright to do doggie style...that way.
Theoretically? Probably.
But the only way to be absolutely sure you are catching all of the violations is to store enough of the work itself that your filter can recognize it when it sees it.
A hash would work, for a specific version of the copyrighted work. However, I could add one sentence to the beginning of the document and the hash would then be different and the filter check would fail.
The only way to be sure is to somehow store enough information about the actual text to recognize that specific text each time you see it. I suppose you could hash every other paragraph and look for matches to the paragraphs, and for the written word that would probably work OK.
But Scribd obviously thought, and the author's lawyers have reluctantly accepted, that storing the original document to have something to compare to is a reasonable method of filtering.
Remember, Scribd is not accused of distributing the copy they retained in their filter. They retained it for the purpose of preventing distribution. And it seems to me that, by doing so, they came up with an extremely effective filter that gave the original work far more protection than is required under the DMCA.
If anything, Scribd should be commended for going above and beyond the letter of the law in order to protect copyrighted works.
"This post contains words, known to the State of California to cause thought. Wash brain thoroughly after reading."
If they found them guilty, then it would have been impossible for them to filter copyrighted stuff, making many people's lives easier. "Sorry, we can't do copyright filtering without violating copyrights..."
>Stealing a car to look for stolen cars doesn't make you a cop.
Making false comparisons between copyable abundant data and non-abundant physical goods does not make you a good analogist.
Corporation, n. An ingenious device for obtaining individual profit without individual responsibility. - Ambrose Bierce
Seriously? Scribd was essentially providing a service saving the author money, and that's grounds for a LAWSUIT? So to ensure that they have nothing illegal on their servers Scribd (and others) should be bound to purchase a copy of EVERY BOOK?!?
Side A: You are bound by law to prevent that from being copied.
Side B: What is "that"?
Side A: Give us money and we'll allow you to know what your bound by law to prevent doing.
IANAL but that smells of extortion to me.
By the way I think your argument using cars should win an award... but not in a good way.
Abundance has ZERO to do with ownership.
Stealing a blade of grass off my lawn makes you a thief.
Got it?
Now, get off it.
Scribd possessed a copy of a book that they did not pay for, and using it as part of a software program.
That's grounds for a lawsuit.
Whether it "saved" the author money is moot. They could have saved the author money by manually examining uploaded content, instead of by using a pirated copy in their software.
Thanks for reaffirming what's wrong with your legal system then.
Your understanding of "abundance" is lacking.
When I breathe, I am not stealing your air.
When I post on Slashdot, I am not stealing your bandwidth.
If every blade of grass remains on your lawn, what was stolen?
Captcha: "Imbecile" No shit.
Infringed from Orwell's work. The story 1984 has that as the central theme, concentrating on the fact that it is partial enforcement that is used to keep people down, NOT merely lots of laws.
After all Randians would NOT like the bit about "don't enforce the laws against the rich and connected". Hence her gutting of 1984's main thread.
Wow. You're justifying a legal argument which not only upholds copyright as a justified monopoly over the use of a work (I can agree with this, mostly) but also forces businesses to buy copies of that work in order to operate a law-abiding service on the supposition that illegal behavior might otherwise occur.
Basically, you're insane or evil.
This could be a useful precedent for the music business. A growing problem is: You're a musician who thinks you've written a new tune, but think you might be pulling a George Harrison and merely putting new words to someone else's tune. How do you find out if the tune in your head is copyrighted?
Some significant work has been done with putting music online in several computer-encoded forms, with several lookup tools that try to identify similar pieces of music. Leaving aside the fact that this is a nontrivial problem musically, the work has been seriously held back by fears of the copyright enforcers. There have been demands from publishers and agents that copyrighted music be removed from the databases, which has usually been done. So the available music is pretty much just material that's out of copyright. It's obvious that you can put Bach's and Beethoven's music online in computerized form, and it's equally obvious that "folk" tunes published centuries ago (by Playford et al) is legal.
But I've actually asked some music publishers' reps how one might legally find out if they hold a copyright for a particular tune. Their answer, apparently told with a straight face, is that I should buy a copy of everything they've ever published, and search through all of it for the tune. There is a certain lack of practicality to the idea that every musician should own a printed copy of every piece of music that's under copyright. I'd need a house orders of magnitude larger than my current house to hold it all. And the search operation would take several years of full-time work.
So what's the prospect for this decision being applicable to music? Would it be legal for musicians to transcribe lots of copyrighted tunes, putting what's essentially a computerized "fake-book" version of everything in our database, and make the lookup tools available online to musicians? Or would we be risking a multi-million-dollar years-long lawsuit over such a tool? Asking publishers for permission isn't feasible; many of them have already said they don't approve.
If it's not legal, how might a musician learn whether that tune in their head is copyrighted? You can play it at a gig, and wait for the indictment papers to arrive at your door, of course, but it'd be nice if there were some better way that's legal. It'd be even nicer if the database could have contact info for each work, so the musicians can quickly ask for permission to play a work in public. But at present, such a tool seems illegal, on the ground that its database is inherently a mass of copyright infringements.
(There's also the separate problem: How to decide whether two tunes or songs are "the same music" is an unsolved problem, whose only answer right now is "Ask the courts", and it's not obvious how to write a CGI program that does that. ;-)
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
Abundance has ZERO to do with ownership.
Stealing a blade of grass off my lawn makes you a thief.
Got it?
Now, get off it.
And if someone steals a blade of grass from your lawn, they should be required to recompense you for the value of that blade of grass. Which, due to abundance, is basically nothing.
Get it now?
Don't tell me to get a life. I had one once. It sucked.
I don't own the air. You're free to breathe it in. Just don't breathe it out at me if you have a disease.
When I access slashdot I grant you a license to the bandwidth needed to read your post.
If every blade of grass remains on your lawn, what was stolen?
If you haven't taken one, you haven't stolen one.
Captcha: "Imbecile" No shit.
There's a reason you refused to sign your post.
No, I am saying that if they choose to use a copy of the work as a means of automatically ensuring compliance with the law, then they have to pay for that copy.
There are other means to comply with the law. They don't have to break the law in one way to comply with it in another.
What if you're training filtering applications to automatically discover and block child porn? ...
I heard a story about a similar problem with a clever solution. According to the story, IBM wanted to eliminate all "bad words" from all of IBM's computer programs. The expectation, clearly, was that there might be bad words in the comments, which would offend IBM's customers. IBM has customers all over the world, and they read a large number of languages, so the directive was to eliminate all bad words in all of the languages used by IBM's customers.
The programmer given this task was not fluent in all of the languages used, so he asked IBM's field offices in each country to provide him with a list of the bad words in the local language. The field offices resisted, because even writing down the bad words offended them. Even if this were not a problem, there was the issue that the resulting program, and its data, would be IBM software, and therefore subject to itself.
The solution, according to the story, was to have the field offices send the bad words spelled backwards. This removed their offensive nature, and also solved the second problem. The program reversed each line it read, and then compared with all of the reversed bad words. The result: a program with no bad words that can detect the presence of bad words.
Application of this trick to the copyright problem is not so simple. A copyrighted work, reversed letter-for-letter, is a derived work of the original. Using ROT13 would have a similar problem. Child porn, on the other hand, isn't child porn any more if the image is not recognizable. You could do something like stretch the image horizontally until it becomes visually meaningless. Do the same for each input image and compare with the "neutered" child porn.
Grass is not a book. It took me a week to grow that grass from seed. Seed I dind't have to invent. Seed I bought. Regardless, it is my property, and stealing it makes you a thief.
If I did invent the seed, and it was valuable to someone, and you stole one of the seeds from me, you bet your thieving ass I'd be coming at you with a phalanx of lawyers and cops, especially if the law says that my failure to protect my rights to the seed causes my rights to lapse.
Copyright is a valid right. Stealing a book, no matter your excuse, and no matter how many trillions of copies someone can produce on a computer in a few seconds, is not a valid right.
Get it now?
The only thing "wrong" with it is that fools like you don't understand it.
Step 1: Private profit-making business wants to make money by hosting other people's work (not necessarily without authorisation).
Step 2: Some woman points out that her work was up without her authorisation.
Step 3: Private profit-making business takes down her work and instead makes a copy of her work as part of an algorithm to prevent further infringement.
Step 4: Woman points out the business is still making an unauthorised copy of her work.
Step 5: Slashdot protects the "right" ("sense of entitlement") of the private profit-making business to make a profit, arguing that it would be somehow "unreasonable" or "evil" for exceptions to copyright law to be made to private profit-making businesses which would otherwise find it "difficult" to operate.
Essentially, /. is pulling an MPAA/RIAA. It's not this woman's problem that it's difficult for Scribd to otherwise police for copyright infringement. It could keep short extracts; it could keep summary statistics (not just a hash - any amount of data on sentence/word structure/length which wouldn't be significantly adjusted by making small tweaks); it could hire humans. It could even find that its business model is unworkable. But no - /. thinks, like MPAA/RIAA, that some corporation should have the right to its profits, and that exceptions should be made in law to protect the corporation's sense of entitlement to profits.
Even ignoring the practical impossibility of doing this, presumably they would still have had to purchase a copy of that, and every other, book (and probably one copy for each person that would be tasked with checking the documents) in order to do it legally.
Then all is not lost and that business model might still have a bright future.
All we need to do is include books in shrink wraps, and have an EULA you must agree to before openingg, saying you may not use the book in a copyright infringement prevention system.
I need to start a filtering service and make the data available to everybody else so that they can start their own filtering service.