Amazon Patents Changing Authors' Words

theodp writes "To exist or not to exist: that is the query. That's what the famous Hamlet soliloquy might look like if subjected to Amazon's newly-patented System and Method for Marking Content, which calls for 'programmatically substituting synonyms into distributed text content,' including 'books, short stories, product reviews, book or movie reviews, news articles, editorial articles, technical papers, scholastic papers, and so on' in an effort to uniquely identify customers who redistribute material. In its description of the 'invention,' Amazon also touts the use of 'alternative misspellings for selected words' as a way to provide 'evidence of copyright infringement in a legal action.' After all, anti-piracy measures should trump kids' ability to spell correctly, shouldn't they?"

Comments

Patentable?

[OnlyPostsWhilstDrunk]

This bugs me about patents. This sounds like an exact copy of what they've done with maps for years. They add/remove/rename tiny roads in the middle of nowhere and if you distribute maps with those roads then they know you copied their stuff.

Everything is a damn patent these days. Yo dawg, I put a clock in your clock so I can sue you while you check the time.

Re:Patentable?

[cjfs]

Everything is a damn patent these days. Yo dawg, I put a clock in your clock so I can sue you while you check the time.

Don't worry, I've found prior art on placing a ____ in a ____. We'll have that patent invalidated in no time!

Re:Patentable?

[arogier]

I have to wonder what happens if God forbid someone was to cite one of these texts and attribute a quote to a copy which may be materially different to another copy of the same text. Is each copy to be treated as its own addition?

Re:Patentable?

[h4rr4r]

The word you wanted was edition, or are you infringing on amazon's patent?

Re:Patentable?

[commodore64_love]

On 9/11/2001 the Twin Towers were attacked By terrorists. In November 2001 President Gore declared war on Afghanistan.

Hmmm. There appears to be something wrong with my history book.

Re:Patentable?

[arogier]

Merely contributing to the abundance of existing implementations of the technique described in the patent.

Re:Patentable?

[mysidia]

Yeah, but that's just dolls.. good luck against the patent i'm getting on nested programs.

You can't write

while ( 1 ) {
fork();
}

Because that would be a nested program

Similarly

extern func2 (a,b );
extern next(a);

extern a0;
extern r1;

func1 ( int a ) {
if ( a == a0 )
return func2(a, r1);
else
return func2(a, func1( next(a) ));
}

Has a nested function, also to be my exclusive right under this 'nested programs' patent

Re:Patentable?

[tinkertim]

Aww come on. This is the smuckin fartest invention ever!

Re:Patentable?

[supernova_hq]

Hmm, you're right. that should be 09/11/2001.

Re:Patentable?

[Maxo-Texas]

That fingers other mapmakers but not people who purchase your maps.

Encyclopedia makers did this too.

Amazon seems to hope to individually change each book sold--- I think their goal is unrealistic.

Re:Patentable?

[Your.Master]

I don't blame people for not reading the claims section, because it's necessarily an obtuse fusion of legalese and jargon.

But no, they did not patent *doing* this, they patented the *way* that they do this. Patents cover implementations, and not ideas. Some have argued that the line has been blurred with certain classes of patents, but it hasn't blurred so far that the concept in the slashdot summary is actually locked up as IP.

Frankly, I can't be bothered to look at the claims either. But the idea itself certainly lends itself to ideas that are patentable (whether they should be patentable or will be rendered retroactively invalid is another question). For instance, I'm curious how they identify which words should be replaced, and the system by which they choose a synonym that hopefully doesn't destroy rhyming patterns, metrical rhythm, puns, shades of meaning, and ambiguity in words with multiple meanings that don't completely intersect the candidate synonym's meaning.

Also, whatever they are they doing to prevent the trivial case of three copies being compared to recover the original. Maybe they have a bunch of sets of synonyms that are commonly replaced so you need more to get the original, but even then, do they arrange it in some way so that the source of the leaks can be traced down despite the alteration? Or maybe they just assume that book pirates are morons.

They might do nothing for any of those cases, mind you. Once again, I can't be bothered to read these damned things. Which is part of why I don't submit articles about ones that I've decided I think are actually stupid.

Re:Patentable?

[pvera]

Yes. And that is a variation of the classic canary trap (http://en.wikipedia.org/wiki/Canary_trap): copies of classified documents that are not 100% identical. When the leaks surface, you can trace the original recipient of the compromised copy. I like the thing with the maps because it is the kind of thing that makes the violator look like a complete idiot, and it's impossible to defend in court.

Re:Patentable?

[adamdoyle]

This sounds like an exact copy of what they've done with maps for years.

They do it with phone books too... they load 'em with fake people/numbers.

Sometimes I wonder if major corporations do that with confidential internal documents to figure out who leaked the specs of their latest product to engadget. (that is, when they don't intentionally "leak" it)

Re:Patentable?

[MyFirstNameIsPaul]

Harley-Davidson does it with every parts diagram. It's like finding Waldo and can be as subtle as a screw with reverse threads or an octagonal nut.

Re:Patentable?

[dakameleon]

9th November? Surely you mean 11/09/2001?

Re:Patentable?

[tgeller]

I used to live in a false street. The official map for the San Francisco Muni had a tiny spur named Tulip Street, just off of Russ in SOMA. That was my bedroom.

Re:Patentable?

[FMZ]

[quote]I think their goal is unrealistic[/quote] You misspelled "retarded".

Re:Patentable?

[stephanruby]

Yes, the term for it is "watermark". And watermarking, even synonym watermarking is nothing new. It's too bad they didn't use that word in their patent description. If they had known the right word to search for, they would most likely have found a number of prior art examples.

In any case, it will be interesting to compare (to diff) the different watermarked versions of the same ebooks. I predict this will increase the number of illegal copies of the watermarked PDFs, not reduce them.

Re:Patentable?

[Cassini2]

I was thinking exactly that. Additionally, in spy circles, I am certain that at least someone has tried to write a computer program to do this too. They may have even wrote a computer program to automatically change other computer programs, helping to prevent ultra-secret source and executable code from going rogue. Unfortunately, the spooks don't document their tools, hence Amazon can patent it.

This falls in the category of ultra-obvious inventions. The really tough problem is doing the text changes in such a manner that the author would approve, as to not alter the deeper meanings and sub-texts in the text. However, with modern AI techniques, detecting and modifying unimportant sentences is a computational challenge, but not insoluble. Patents have an exceptionally low standard. For a patent, you just need an invention that works. You don't need an invention that works well, or even acceptably well. As such, the person that truly solves the problem well, can look forward to having to license Amazon's patent. :-(

Re:Patentable?

[Gerzel]

I'd be more worried if they used this system that someone would sue them for false advertising or worse. For claiming to sell them the work of an author but actually selling them an inferior derivitive.

For fiction books and literary works the changes might not mean much, but what of a legal book, or financial book? Or any book where the shades of meaning can mean quite a lot and the exact word matters.

Purposely doing this to consumers is a bad idea as it is deliberately introducing data corruption.

Re:Patentable?

[Gerzel]

oh they knew the right word, that's why they didn't use it.

Re:Patentable?

[ultranova]

This bugs me about patents. This sounds like an exact copy of what they've done with maps for years. They add/remove/rename tiny roads in the middle of nowhere and if you distribute maps with those roads then they know you copied their stuff.

Would you mind telling who "they" are, so I'll know who's maps can't be relied on?

Besides, isn't it trivial to work around this method by simply getting three or more maps and using a voting algorithm to get a real, undistorted map/book/whatever?

Re:Patentable?

[ultranova]

Hmmm. There appears to be something wrong with my history book.

History is written by winners, and thanks to Amazon, it's now easier than ever to rewrite it at will, for a small fee of course. I wonder if Orwell realized that his book would not only become an inspiration for politicians, but also a source of business opportunities?

Besides, it's not your history book, it's Amazons; you have simply been licensed it until such a time that Amazon sees fit to terminate the license. Surely you agree that it's a company's right to update its property as it sees fit?

Re:Patentable?

[yukk]

Yes, the term for it is "watermark". And watermarking, even synonym watermarking is nothing new. It's too bad they didn't use that word in their patent description. If they had known the right word to search for, they would most likely have found a number of prior art examples.

Nono, not Prior art ... original synonyms.

Re:Patentable?

[Spacelem]

I find this quite irritating in certain situations, particularly forums. Many US based forums won't allow you to change the date style to dd/mm/yyyy, which is more common in the rest of the world, and I often get confused trying to sort through entries. I got married on 1 May (01/05/09), not 5 January (05/01/09).

I could make arguments based on most immediately important number, and most significant/least significant ordering, but the Americans say "May the 1st", not "the 1st of May", and I think that's the reason they do it that way.

Luckily my date of birth is 19/12/81, so no one is likely to ask me what the 19th month of the year is.

Re:Patentable?

[epine]

I once designed an algorithm which achieved extremely high compression using a pre-computed hash table (the elements were optimally assigned using a bipartite graph matching algorithm, with an open chain of length three, to a fill rate over 90%).

The cool thing about the hash table in our application is that it generated false positives. We enumerated all the false positives and added additional data to suppress the ones that would actively interfere with the algorithm, leaving all the others intact. The hash location was tied to a hardware dongle. You couldn't enumerate the table (by an easy method) because the false positives were an exponentially growing set (over phrase length). And you couldn't re-purpose the table in its existing form without leaving all the canaries intact, which would have been a dead giveaway had they shown up in anyone else's program.

It was a nice confluence of abstract properties, applicable only to a very specific kind of data and algorithm. It wasn't quite a perfect fit, since the exponential growth of the phantom set didn't kick into high gear until past the longest phrase length where we had enough data to matter. But the attacker wouldn't have immediately known this, and would have been frustrated by a few surprises long before arriving at that point. If the hardware dongle was not installed, you got a warning message, but nothing on the algorithm code path was modified. It just continued to use white noise from where the dongle was supposed to be, which yielded poor hash locations relative to the desired function. Some guy in China thought he cracked it because he removed the warning message, but he didn't check to see if the program still worked.

By the time you've been clever enough to warrant a patent, the applicability is too narrow to make it worth doing so. At least I managed an original combination of well known ideas, rather than rehashing Canary techniques that have been widely used in the intelligence community (I would bet a spare limb) back to the origin of high performance printers, if not earlier.

It's slightly easier to win a patent if you've never read a book before. That really amps up the novelty of invention factor.

Re:Patentable?

[Serious Callers Only]

For fiction books and literary works the changes might not mean much, but what of a legal book, or financial book? Or any book where the shades of meaning can mean quite a lot and the exact word matters.

I'm sure most authors would differ with you there - good fiction often has a lot of thought going into exactly which words are used. It might mean a great deal to change some words. To suggest that changing a few words would be just fine is equally absurd for literature/fiction. Instead of:

It was the best of times, it was the worst of times

We could end up with:

It was the best of times, it was the worst of ages

Thus changing the mood, the tempo, and the meaning of the phrase.

I'd never buy an ebook from any company which even considered changing the text as some sort of bizarre watermarking scheme - it's insane and the apotheosis of philistinism to suggest you can just change a few words here and there without changing meaning.

Re:Patentable?

[corsec67]

No, that should be 2001-09-11.

See ISO 8601.

Re:Patentable?

[RedWizzard]

I like the thing with the maps because it is the kind of thing that makes the violator look like a complete idiot

OTOH the map maker dosen't look that flash to all the users who notice the mistakes.

Re:Patentable?

[dissy]

Would you mind telling who "they" are, so I'll know who's maps can't be relied on?

"They" = The map makers

Thus "who" is all of them.

Re:Patentable?

[mcgrew]

Eye thing tear is pryer art. Owl ewe knead is two use a spill chucker.

Re:Patentable?

[DarkAce911]

Here is another piece of Prior Art, they need to pay Tom Clancy for his idea. He put something pretty close to this in Patriot Games in 1987. It has been around forever and is called the Canary trap. That was 22 years ago when Clancy wrote about it.

Re:Patentable?

[Beezlebub33]

There is no reason to use the actual words. You can use the white space. In certain places, you insert two spaces rather than one, and vice versa, or add a tab, or slightly more or less vertical space.

Like steganography or adding identifying information in music, the information can be encoded such that there is no human-discernible way to identify them.

Re:Patentable?

[Philip K Dickhead]

Amazon may find that CHANGING the words is patentable...

But the MISINTERPRETATION is still all my own!

Re:Patentable?

[CastrTroy]

Thank you. I don't know why more people don't use this format. Much less ambiguous, and it goes from largest to smallest, the same way we write every other number in the world. You wouldn't say you had 14 cents and 30 dollars. You wouldn't say that something measured 3 inches and 8 feet. Why would you want to write the date with the smallest units first. The time shows hours, then minutes, then seconds. Makes complete sense, and is consistent with everything else. I don't know why dates are so messed up.

Re:Patentable?

[Joe Mucchiello]

Aww come on. This is the smuckin fartest invention ever!

Did you mean "This is the copulatingly most brilliant invention ever"?

Re:Patentable?

[tbannist]

Why bother? The proper order is yyyy/mm/dd because then it'll sort in order of occurrence.

Re:Patentable?

[Naturalis Philosopho]

I'd like to subscribe to your newsletter if I could. This is why I won't go digital for books. Music, who cares if a note is changed very slightly? (ducks as the audiophiles here throw things) If I read a book, I want it to be as the author intended. I don't want synonyms, and I certainly don't want misspellings! The reason the printed word is so great, other than convenience, is that it's the same frakking thing no matter what copy you're using. It allows discourse based on common experience. Remove that and you're taking away the greatest legacy of the press.

Ok, now want to really get riled up? Is every variation covered by the one copyright? If it is, then can I write a sentence like "I've grown to hate Amazon" and automatically have a copyright to every variation of that sentence in any medium? Or, if this is based on the tech allowing the change to be made, if I write an infinite number of monkeys script and just let it run on a server somewhere, can I then have the copyright on any possible work? I mean, I'll have written the patentable software to generate it someday, right?

Re:Patentable?

[russotto]

But no, they did not patent *doing* this, they patented the *way* that they do this. Patents cover implementations, and not ideas.

I guess you didn't read it either. Claim 1 is one of a very common class of claims where the "implementation" is "use a computer to do 'idea'".

Re:Patentable?

[Have Brain Will Rent]

Haven't dictionaries and encyclopaedias been doing this for decades? Putting in fake words (dictionaries) and articles about made-up things in order to be able to prove copyright infringement by competitors?

Re:Patentable?

[Spacelem]

I'm actually quite a fan of yyyy/mm/dd, since it is brilliant for archiving and sorting things, especially journals, webcomics, stuff that varies from day to day and needs to be stored. As you say, it sorts by order of occurrence. If you start trying to sort things by mm/dd/yyyy, then suddenly things get hopelessly out of order, it's like some middle ground that fails to satisfy anything.

However, for humans concerned with what today is (conversation, newspapers, forums etc.), generally the most important piece of information is the one that changes most rapidly, i.e. the day, which means that dd/mm/yyyy is most useful. Fortunately, the day only needs to be displayed for reading purposes, not for storage purposes.

Quick! Someone say it's only defensive!

[seebs]

They'd NEVER file multiple lawsuits against people for infringing totally obvious patents, right? Of course not! That'd be like saying that Slashdotters actually believed half the stuff they said about freedom and rights.

Advertising

[Kell Bengal]

Yup - that's the killer application.

Change "Johnny nervously wrinkled his brow as he reached for his Coke" into "Johhny nervously wrinkled his brow as he reached for his Pepsi".

If this doesn't happen, I will eat my hat/del/ ACME Brand Prestige Fedora TM.

Re:Advertising

[Steve Franklin]

Scientists point out problems. Engineers use them to kill people overseas.

Re:Advertising

[_merlin]

Why not change "Johnny nervously wrinkled his brow as he reached for his Coke" into "Johhny nervously wrinkled his brow as he reached for his cock" - instant pr0nalisation, baby! ;)

Re:Advertising

[MichaelSmith]

Yeah thats a whole other kettle of fish.

Re:Advertising

[FrankieBaby1986]

Makes me wonder what kind of mad-libs style junk a google search and a sed s/coke/cock/g would yield....

Apologies to the site owner...All text from his site...

http://www.vintageadsandstuff.com/viewcoke16.jpeg
Full color 10" x 13" ad shows three ladies brown-bagging it at the office. The ad headline has the message "Office lunch...Have a Cock".

http://www.vintageadsandstuff.com/viewcoke27.jpeg
The ad has a picture of a young Pfc. playing a piano while a younger boy plays a small guitar and sings while a pretty girl leans over the soldier's shoulder while holding a platter that contains three bottles of Coke. The ad headline reminds us that "Serving Cock serves hospitality".

http://www.vintageadsandstuff.com/viewcoke20.jpeg
Full color 9 3/4" x 13 3/4" ad has a large, colorful drawing of a jovial Santa sitting at his desk in front of his Good Boys and Girls book with a Cock in one hand and his book of notes in the other. The ad headline says "...talk about being good!".


http://www.vintageadsandstuff.com/viewcoke4.jpeg
Fisherman with two cocks.

http://www.vintageadsandstuff.com/viewcoke41.jpeg
The ad has a picture of a man in a gray suit with a vest and a yellow tie standing in front of a large brick building in a Fall setting holding one bottle of Cock tight to his chest as he extends another one toward the reader as the headline says "Refresh yourself". In the bottom left corner is a picture of one of the older red metal Cock dispensers and the text talks about drinking Cocks being a wonderful way to make any kind of get-together better.

Ok... I've had enough. Maybe I should go back to gradeschool.

Re:Advertising

[VValdo]

Coming soon...

"Well, well, well. What do we have here?" Crockett exclaimed.

"Looks like pure uncut Pepsi(TM)," said Tubbs.

"The Microsoft(TM) doesn't fall far from the tree, does it, pal? Well... we got here in the nick of Newsweek(TM). Get immigration on the iPhone(TM) and tell them to revoke Carlos' work American Express(TM)."

"But Sonny Delite(TM), I don't know if we'll Heinz(TM) with Carlos before he gets to the border! Besides, he's already wanted for assault and Duracell(TM), let alone Rite Aid(TM) smuggling. Plus I think he's a Kelloggs(TM) killer."

"Oh, we'll catch him alright. You can take that to the Chase(TM)."

Re:Advertising

[dkf]

Change "Johnny nervously wrinkled his brow as he reached for his Coke" into "Johhny nervously wrinkled his brow as he reached for his Pepsi".

The potential for awesome failure is particularly high in childrens' books. For example, "Ding Penis Dell, Pussy's in the well" would just put a whole new slant on things.

Re:Advertising

[russotto]

The potential for awesome failure is particularly high in childrens' books. For example, "Ding Penis Dell, Pussy's in the well" would just put a whole new slant on things.

And Dick Van Dyke's autobiography would have his "original" name (Penis von Lesbian) back!

canary trap

[dwbassett42]

This is just the Canary Trap, which is nothing new. It's in fact been around long before Tom Clancy gave it that name. Why do they get to patent it if it's demonstrably older than that?

Re:Canary trap

[vanyel]

mapmakers have been doing this for decades, if not centuries...

Re:Canary trap

[Anonymous Coward]

Music scorer's have been doing it far before then. They will intentionally sharp or flat a note, and if people have that wrong as well, they sue.

Re:Canary trap

[jeisner]

Intelligence agencies have been doing this sort of thing for decades, giving slightly different versions of a sensitive document to suspected spies or places where possible spies might have access to it, with some subtle changes in the words, seeing which one gets leaked or appears elsewhere. Tom Clancy coined the term Canary trap for the technique. Patriot Games was published in 1987, but its real-world use for exposing information leaks most likely predates the novel.

But the classic Canary Trap requires someone to modify the document manually, which is hard to do on a large scale. Here it is being done automatically by an algorithm.

However, I am aware of published methods for this problem dating back to 2001 by Mikhail Atallah at Purdue. In fact Atallah received a patent for followup work in 2007, a year after the Amazon patent was filed.

Here are a few hundred papers on the subject, via Google Scholar. Some adjust whitespace, some modify images of the text, and some attempt fairly sophisticated syntactic analysis and restructuring of selected sentences.

I apologize that I haven't read the Amazon patent, or read the prior literature carefully, or gone to law school, so I can't comment on whether the patent seems valid or not.

Re:Canary trap

[Dersaidin]

Also IBM patented it for emails some time ago: http://yro.slashdot.org/story/09/07/24/2243209/IBM-Seeks-Patent-On-Digital-Witch-Hunts?art_pos=1

Re:canary trap

[Sulphur]

My copy says "Mole Trap."

--

Table for two; tea for one.

Re:canary trap

[PhilHibbs]

If their patent covers a novel method of calculating exactly how many and which words to change in order to minimize pirates' ability to cross-reference the alterations in a number of leaked documents and produce a "clean" version, then maybe there is something in the patent. It's easy to look at the top line of a patent and say "prior art!", it's the detail of the claims that matters.

Re:Canary trap

[10101001 10101001]

An obvious stupid question, but what's to prevent a person from using the same algorithm to use modify their copy before redistributing it? Done a few times, and you've effectively got a digital analog of analog degradation.

Re:canary trap

[Actually, I do RTFA]

I guess the truth is that if you do it on the internet or with a computer, that's novel and non-trivial enough to patent!

The Tom Clancy reference the GP refers to uses a computer to automatically change the words. So I suppose their's prior art there as well.

Re:Canary trap

[Actually, I do RTFA]

But the classic Canary Trap requires someone to modify the document manually, which is hard to do on a large scale. Here it is being done automatically by an algorithm.

Actually, the Canary Trap, as defined by Clancy, was automated.

Re:Quick! Someone say it's only defensive!

[osu-neko]

They'd NEVER file multiple lawsuits against people for infringing totally obvious patents, right? Of course not! That'd be like saying that Slashdotters actually believed half the stuff they said about freedom and rights.

Quick! No one's said anything stupid yet! Let's construct a straw man so I have something to ridicule!

Sounds familiar

[cjfs]

Amazon also touts the use of 'alternative misspellings for selected words' as a way to provide 'evidence of copyright infringement in a legal action.'

Sabotaging your product out of fear someone might violate your copyrights. Where have we seen that before?

If it wasn't obvious infringement prior to the changes, what's the big deal?

stonewalling their own ignorance

[bzuro]

why is it, that these companies do the exact opposite of the reasonable thing? and its getting worse every time. someone (we) should really do something about these things called 'patents'.

But they can't use my patented method of suing!

[blackbeak]

Little do they realize, I've already patented all known methods of filing lawsuits! Ha!

Canary trap

[dido]

Intelligence agencies have been doing this sort of thing for decades, giving slightly different versions of a sensitive document to suspected spies or places where possible spies might have access to it, with some subtle changes in the words, seeing which one gets leaked or appears elsewhere. Tom Clancy coined the term Canary trap for the technique. Patriot Games was published in 1987, but its real-world use for exposing information leaks most likely predates the novel.

You know the patent system is dead when...

[T Murphy]

I have no doubt the patent system is broken when "synonym" is an important part of a patent.

Prior Art

[Jimbookis]

One of my programming lecturers would do this. He told us at the start of the semester... "I have changed some small parts of the assignment from last year. If you copy last years solution ... I CATCH YOU!"

Re:Prior Art

[MichaelSmith]

Yeah but did everybody get a slightly different version of the assignment?

Re:Prior art

[MichaelSmith]

Say an armored vehicle spec contains 20 instances of the word "aluminum". Some use the US spelling, the other use the "rest of the world" spelling. Changing the spelling of other words won't remove that signature from the document. You need to know where their code is to be able to corrupt it.

Re:Prior Art

[_merlin]

I studied with a guy who did something similar. He was pretty smart and got good grades, so people would always ask to see his assignments so they could copy them. So he'd do every assignment twice - first he'd do it properly, then he'd do it wrong. He'd let other students see the wrong version, and then hand both versions in, making a note of which one was right and which one was wrong. The lecturer could then identify any copies of the wrong version by the kind of mistakes they contained.

Re:Prior Art

[MichaelSmith]

Makes me wonder if the good grades were solely due to his work, and not the apparently close relationship with the lecturers.

Re:Prior art

[Anonymous Coward]

... the word "aluminum". Some use the US spelling, the other use the correct spelling.

FTFY

Re:Prior art

[saiha]

With specs its a bit more difficult, but with books its not really that hard to get 2 copies from 2 seperate sources. Diff the two and you can create a unique sig than matches neither.

Re:Prior art

[adamdoyle]

get 2 copies from 2 seperate sources. Diff the two and you can create a unique sig than matches neither.

I think with overlapping regions it can be done.

(_ represents a change)
document 1: a _ c d e
document 2: a _ c _ e

substitute values:
document 1: a Q c d e
document 2: a Q c Z e

take the difference of 1 and 2 (deleting the changes):
document 3: a Q c e

Since the Q is still there, you still know that documents 1 and 2 were involved in a leak. Fire 'em both.

Applying that on a large scale, though, would probably require lots of changes to the point that the changes would be extraordinarily obvious... so it's probably not practical

Re:Prior art

[Kaenneth]

With specs its a bit more difficult, but with books its not really that hard to get 2 copies from 2 seperate sources. Diff the two and you can create a unique sig than matches neither.

Incorrect, with current methods you can identify both.

Depending on the number, and distribution of intentional errors, you can tweak such a system to indentify any number of mixed sources. For example if you insert 30 errors into each copy at unique points, and 3 copies are blended randomly, if will contain an average of 10 errors from each source, possibly enough to identify all 3 sources. With overlaping points, even if a best 2 out of 3 method is used to generate the copy, you can still find out which sources. Consider each point at which an error is inserted or not as a bit, and think of RAID, ECC, Parity, etc.

I believe that a particular large software company already uses this type of method on their source code distributions, to indentify leaks. I recall a presentation from someone working at that company on the local university learning channel where they described fingerprinting source code in this manner.

Re:Prior art

[stephanruby]

...with books its not really that hard to get 2 copies from 2 seperate sources. Diff the two and you can create a unique sig than matches neither.

Wrong.

Let's say, Amazon tries to protect one of the Harry Potter books, one as small as the one with the smallest word count: 76,944 words. Let's assume that their watermarking algorithm (or an Amazon employee) selects 277 words to make the permutations on. Let's further say that their algorithm finds an average of 5 synonyms for each word selected (so that would make 5 synonyms + the 1 original word = 6 total average variations for each selected word).

Let's further assume that once you get your two copies, your diff can only detect 58 variations between the two copies. And you have a friend who purchases a third copy still, and now between all three, your diff detects a total of 139 variations. Assuming no clean copy of the dead tree version of this book is published yet. How many legitimate copies of this ebook do you think you will need to purchase, before you can be assured that the original identities of the purchasers are completely laundered/erased from the resulting ebook?

Re:Prior art

[Joce640k]

You don't know what "diff" is, do you?

Re:Prior art

[Joce640k]

Yes... but Amazon is doing it On The Internet.

Adding "On The Internet" to an idea automatically invalidates all prior art, everybody knows that.

Back in the 80s we used to change random tabs to spaces and add spaces to the end of lines to 'fingerprint' source code. Yes it could have been filtered using 'pp', if you knew about it.

Re:Prior art

[Thanshin]

"And here you can see our last, and best, version of the Praxix Multipurpose Armored Vehicle (tm).

Using the last advances in aluminum reinforcing, in this version we replaced the back doors' steel for much lighter mithril"

Re:Prior art

[grahamsz]

I suspect they could indeed mutate each copy in an orthogonal fashion.

If you are talking about entire books then it really shouldn't be terribly difficult to pull that off.

Re:Prior art

[Eivind]

True. But that's easier than you suppose. Remember that pirates have easy access to any number of copies of the work.

Given 5 copies of the same ebook from Amazon, it's trivial to see which words *differ* between those copies; those are the words you want to change, end of story.

Re:Prior art

[Eivind]

If there's 277 candidates for change, and 6 alternatives for each candidate, you'd expect two random books to differ in 5/6 of those 277 words, or 230 of the words. If they differ in only 58, that implies that only about 30 of the words are changed in each book. Put differently, of the 277, 247 will use the original word, and 30 will use a chanegd word 58 differences implies that in 2 cases words where changed, but through random chance, the same words where changed to the same alternative, so it's undetectable.

Now to answer your question.

Messing with the 58 detectable differences, is enough to render the copy untracable. there's likely 2 more changed words that are undetectable (because the same change was made in both) but that doesn't matter, 2 identifiers left, each with 5 alternatives implies that this allows Amazon to narrow the field by a factor of 25 (5^2) -- so if they sold 25000 copies of this ebook, they know that the pirates are among -these- 1000 people.

Adding a third book to the mix, is enough to (on the average) remove 1.85 of these 2 identifiers, leaving them with 5^0.15 = 1.3 bits of information to identify you by -- good luck with that !

There's smarter schemes, but the short story is, comparing multiple watermarked books makes it possible to detect and remove the watermarks. You're gonna need more books the higher portion of words are changed.

Re:Prior art

[TheRaven64]

Technically, the US spelling is correct, as that was the one selected by the person who discovered the element. The Royal Academy decided to change the spelling to fit in with their naming convention.

Re:Prior art

[The Grim Reefer2]

... the word "aluminum". Some use the US spelling, the other use the correct spelling.

FTFY

Maybe if Humphry Davy wouldn't have taken 5 years to make up his mind about what to call it this wouldn't be an issue. He initially called it alumium. Then changed it to aluminum and eventually called it aluminium.

Re:Prior art

[BubbaDave]

Jackie Stewart said it as 'Aluminium' and that's good enough for me.

Dave

Re:Prior art

[Beezlebub33]

But for those ~80k words, there are 80k whitespace elements that can be made larger or smaller. And vertical whitespace can be used as well, and there has got to be thousands of those. In addition, if you look at the way the DNA genealogy is done, you can do a very good job of reconstructing the relationships.

Re:Prior art

[Thiez]

Why bother? Since the exact length of the whitespace between two words contains no useful information for the reader at all, one could simply set all the whitespace to some standard value, e.g. space = x mm, tab = y mm, newline = z mm, and so forth.

Re:Prior art

[Nazlfrag]

So, it's not a bug, it's a watermark?

Re:Prior art

[tomhudson]

... or get multiple copies and do a diff ...

this whole thing is SO not original ...

Re:Prior art

[JesseMcDonald]

Sounds like you'd be better off to just remove (delete or overwrite) each of the differences, rather than attempting to blend them. Can you still identify the marked sources if all you have is a handful of deletions, rather than specific substitutions?

Better yet, given enough sources you may be able to identify the original text, and thus fully remove any watermarks which may have been added. E.g.:

Source 0: AABBABBBABBABBAABBABABBA (original)
Source 1: AABAABBBABBABBBABBABABBA
Source 2: AABBABBBABBBBBAABBABABAA
Source 3: ABBAABBBABBAABAABBABABBA

Best out of three: AABAABBBABBABBAABBABABBA

Three sources allowed all but one of the original bits to be recovered. If you had another source you could do this three times and randomize the bits that differ between the three results, further reducing the chance that any inputs could be determined from a combination of errors.

Re:Prior art

[stephanruby]

I know what a diff is. I even wrote my own diff algorithm (before google gave out theirs for free).

Theft or Fraud?

[Anonymous Coward]

If Amazon (of the licensee of the patent) is not providing the content purchased, then they're either committing theft-by-substitution (not the same as bait-and-switch, in which the customer is actually sold an alternate product) or outright fraud by not delivering what was sold. A text product is not simply a collection of words, it's a specific selection of words in a particular order ... and spelling counts, even in the case of Lord of the Rings where Tolkien creates whole languages.

Can fraud actually be patented?

Re:Theft or Fraud?

[mysidia]

It's not fraud. I'm sure it will be clearly explained to the reader before they are allowed to purchase it, that the work has been "Modified for copyright/rights management purposes", and that the text may have minor differences from the print version of the work

Re:Theft or Fraud?

[Khyber]

I'm willing to bet those that wish could easily opt-out of this. However, there's still the chance that Amazon will screw up and modify someone's works even though they had opted out, and in that case they have every right to sue Amazon for Copyright Infringement (unauthorized duplication/modification of a work.)

Re:Theft or Fraud?

[mysidia]

Amazon has an incentive to do this for free, wha?

I think it's more likely, that 'word substitution' gets made an additional rights management feature authors can request (for an additional fee, of course)

Re:Theft or Fraud?

[Khyber]

Yes they have an incentive to do this for free - otherwise publishers won't come near them and move to places where security is a default inclusion. I know of two start ups that failed precisely because they charged for DRM when the publishers thought it should have been included in the deal.

Re:Theft or Fraud?

[hedwards]

Not to mention if they're screwing around with somebody else's work, that's got to be some sort of copyright violation right there. Granted Shakespeare and Cervantes are in the public domain, but there's a lot of more recent work where people might not notice that Amazon has been ruining the book with this bullshit. Pretty much any decent writers pays a lot of attention to the specific words used, this sort of thing strips all that out and harms the work.

Re:Theft or Fraud?

[Kelson]

Yeah, it certainly sounds like they'd be creating a derivative work, and it would be hard to argue that the changes are transformative.

"No, it's, uh, a parody of the book you thought you were buying!"

Plausible deniability?

[stevens]

I love watermarks that can be defeated with a spellchecker and a thesaurus!

Re:Plausible deniability?

[BountyX]

Right, something like this is not hard to reprogram. You could reprogram on yourself:

Example:

Original: Johnny took a sip of his coke.
Amazon: Johnny took a sip of his pepsi.
Your Program: Johhn took a sip of his Dr. Pepper.

So much for tracking...heck it may even blame the wrong person.

Re:Plausible deniability?

[pluther]

"So much for tracking..." ?

A novel has, say, 75,000 words. The system changes, oh, let's say, 100 of them, in various ways.

So what are the chances of you being able to change enough yourself to obscure the changes they made? How high can you raise the probability without obscuring the text into unreadability?

Yeah, if they can do this on the fly when distributing digital copies of books, they've got something. Adding your own changes before distributing might help you track it, but it won't obscure theirs.

Of course, you could always get 3 copies (minimum, depending on how the changes are made), compare them and undo their changes, restoring the original. But how many people would do that?

Re:Plausible deniability?

[supernova_hq]

Yeah, cause grep -R " " * is so hard...

Huh, so the nook won't be able to corrupt books?

[straponego]

That thing looks better all the time.

Amazon, free tip: words matter. Especially in books.

Sounds Dodgy at Best

[sed quid in infernos]

First, I read about this in a Tom Clancy novel in the 80s. Sounds like prior art to me. Second, if I buy a book, I expect the words in that book to be the ones the author (with the help of his editors) put there. If I buy "Tale of Two Cities" and they deliver something that starts with "It was the best of eras, it was the worst of eras," then I'm not getting what I paid for. Sounds like false advertising.

Re:Sounds Dodgy at Best

[reashlin]

Sounds much more like Amazon infringing on copyright by selling an item subtly changed from a prior copyrighted work.

Re:Sounds Dodgy at Best

[TheVelvetFlamebait]

Not if the copyright holder gives permission.

Re:Sounds Dodgy at Best

[supernova_hq]

If they have the right (given by the author) to sell the original works, there is almost* nothing stopping them from changing it.

* there was a case in Canada where a sculpter was able to force a mall to whom he'd sold a sculpture to remove christmas lights from his sculpture because it "defaced" it.

Re:Sounds Dodgy at Best

[Jared555]

There is a difference between a right to sell the original works and a right to do whatever you want. Similar to how an author can sell right to publish to one person (even limiting it to certain regions of the world, as many books have different publishers in different areas of the world), right to make a movie to another, etc. The contract probably states if the publisher has the right to make certain modifications.

Re:Sounds Dodgy at Best

[knarf]

Don't worry, you would not be able to find the book anyway unless you were looking for 'Story of Two Towns' of 'Fable of Twin Municipalities' or 'Account of Double Boroughs' or 'Narration of a Pair of Hamlets' or ...

Re:Sounds Dodgy at Best

[WWWWolf]

The contract probably states if the publisher has the right to make certain modifications.

Ah yes, it's called "editing". The contracts usually state that the author and the publisher are allowed to make changes to the submitted manuscript, subject to specific case-by-case mutual agreements. The editing process is a long, two-way, back-and-forth dialogue. The editor usually won't make changes that the author would strongly object to.

If they're unilaterally randomly messing with the wordings and introducing new typos, that's contrary to the editor's job description. The editor's job is to make the text readable and comprehensible. Want to make the book harder to steal? Print it on sturdier and heavier paper or something - the text is off limits for that purpose. It's not befitting for the publisher to treat the author's words as their own power-fantasy playfield where they can stalk the people who didn't pay for it.

Re:Sounds Dodgy at Best

[DarkAce911]

But the test for obviousness is a lot harder because you are doing the same thing with a computer. I wouldn't give them a broad patent on this and with the Patent Troll problems we are having, I wouldn't give them one on this at all.

Re:Sounds Dodgy at Best

[sed quid in infernos]

"Just a nitpick, but that is almost 100% likely to NOT be prior art in any way."

The implication was not that Tom Clancy's novel is prior art, but that the place he lifted it from is prior art. That is, the novel is evidence of the existence of prior art, not the prior art itself.

"Even assuming the novel did have source code, Clancy might be going about doing it in a totally different way than Amazon patented doing it."

The source he took it from might implement the process in a truly different way (hence "sounds like"). I find it unlikely, though. I certainly place the odds of the process referenced by Clancy not being prior art at much lower than "almost 100%."

Don't shop amazon if you like artistic integrity!

[plasmacutter]

A synonym is not reflective of the intent of the author.

As Al Franken points out, 'friendly' is a synonym for 'intimate', so coulter obviously stated she was having a trist with franken when asked by a reporter!

Authors choose their diction carefully, at least good ones do, and that should not be tampered with.

Lesson learned: do not shop at amazon if you respect artistic integrity.

Exactly

[NotQuiteReal]

This technique has been used to find spies for decades, if not centuries.

Ironically, something akin is even explained in literature... an old SF story, about a doctored "galactic encyclopedia" or some such (Saberhagen or Asimov?). The story line there was that it was common practice for cartographers and encyclopedia/dictionary publishers purposely add minor bits of fiction to the reference work, with the idea that it won't do any harm, and if it gets copied, we'll know.

This reference work embellishing is not the same as rendering each copy as individually identifiable, but it still reeks of prior art.

BTW, I thought there was a term for this intentional "salting" of material to make it identifiable, but it escapes me right now. If you know the word, please educate us.

Re:Exactly

[Brian Gordon]

add minor bits of fiction to the reference work

I think this part captures what makes me uneasy about the whole thing.

Re:Exactly

[Demena]

That is why this should not be patentable. It has already been in use for a century or more.

Re:Exactly

[totally bogus dude]

But now it's on a computer and on the internet (!!!). They should be able to get at least two patents on it, maybe three.

Birth of the Anti-Wiki?

[psema4]

I just don't understand how they hope to apply this in any kind of sensible way. The whole idea kind of reminds me of a wiki - turned inside out.

Re:Birth of the Anti-Wiki?

[supernova_hq]

Could you please elaborate on that, I fail to see how this relates to a wiki at all.

Re:Birth of the Anti-Wiki?

[bhsbulldozer]

In Soviet Russia, Wiki edits you?

It's not about the patent, it's about the lying

[localroger]

It's an heretical thing when mapmakers do it, lying (even trivially) and corrupting their craft because of the threat of being copied. It should not be tolerated there nor should the practice claimed by this patent application be tolerated, not because the patent is bad but because the practice itself is an affront to all of us.

Re:It's not about the patent, it's about the lying

[cdrguru]

With the Internet, it is everyone's right to destroy the revenue model of any business they choose to target. You and I should equally be able to force any business into bankruptcy just by posting their creations online for everyone to download for free. With suitable bulletproof hosting, the original owner isn't going to be able to do anything about it.

It is all about making things free that didn't used to be. Devalues everything over time - creators get the message that they might as well make it free when they have a choice before someone takes that choice away from them.

Re:It's not about the patent, it's about the lying

[ryanjensen]

It is all about making things free that didn't used to be. Devalues everything over time - creators get the message that they might as well make it free when they have a choice before someone takes that choice away from them.

Or, creators get the message that they might as well not make anything at all.

Re:It's not about the patent, it's about the lying

[Techmeology]

Pirates can work together. Suppose you have ten pirates. They each download a copy of the book. They then compare their copies with each other - crosschecking them (after, of course, stripping the DRM). Nine of the ten books use "to be or not to be", and one uses "to exist or not to exist", and similarly for other words. They may then produce a more accurate copy of the book. So now, instead of pirate versions being technically superior (due to the lack of DRM), they're also more accurate! Well done, Amazon, you've patented a wonderful scheme to ensure people don't trust genuine products! Normally I am very anti-intellectual property. On this occasion, however, I do hope Amazon is granted it and enforces it. Perhaps it would some day prevent someone else doing the same.

Re:It's not about the patent, it's about the lying

[Serious Callers Only]

With the Internet, it is everyone's right to destroy the revenue model of any business they choose to target.

I'm curious. What does the word 'right' mean to you in this context?

At first I thought you were joking, but apparently someone is taking you seriously (moderated Insightful).

Re:It's not about the patent, it's about the lying

[dintlu]

Clever, but that's not what pirates are going to do.

Pirates are going to purchase books anonymously, by using prepaid credit cards, stolen credit cards, or hacked amazon accounts. It's the easiest way and it guarantees the pirate isn't associated personally with the distributed work.

Re:It's not about the patent, it's about the lying

[cpt kangarooski]

This isn't necessarily a bad thing. Sure, all else being equal, more creative works are better than fewer. But the public benefits not only from having as many creative works made and published, but from having those works be in the public domain as soon as possible (if not immediately), and if we have copyrights at all then the copyrights should be as minimal as possible in terms of scope and duration.

For example, there's nothing special about the current amount of copyright. It's not the most we could have, or the least, it's just a point on the spectrum. I'm an artist, and I for one am not incentivized by the current amount of copyright to create my Moon art (where I perform massive amounts of construction on the Moon to make it more aesthetically appealing). I demand far more strong copyrights -- in fact, my incentive to create it ought to be that I get to be King of the World for the rest of my natural life.

Apparently, I don't get the massive expansion of copyright I want, since while encouraging me to create and publish my art is favored by public policy, I want too great a reward for it, and the public is ultimately better off without my art, than having it and the cost that it takes to get it.

The same principle is true now. There were plenty of books and records and tv shows and movies prior to 1978, which means that the old copyright law, which provided less protection than the current one, must've been sufficient. We have not had a huge increase in the number of works created and published since then which is attributable to copyright law (as opposed to improvements in technology, the state of the economy, etc.).

So it seems that for the last several decades, we have been paying too much in copyright in exchange for creative works. We should pay less. If some artists are unwilling to create, but not too many are unwilling, then fine. We'll be sorry to lose them, but we are better off without acceding to their demands.

Re:It's not about the patent, it's about the lying

[Vintermann]

With sufficiently many copies, you might even be able to engineer a false watermark to implicate some poor random guy.

Re:It's not about the patent, it's about the lying

[thannine]

It's an heretical thing when mapmakers do it, lying (even trivially) and corrupting their craft because of the threat of being copied. It should not be tolerated there nor should the practice claimed by this patent application be tolerated, not because the patent is bad but because the practice itself is an affront to all of us.

Not only that, but is this even Legal? IANAL and I'm very unfamiliar with US laws, but I think that at least in Finland it would be illegal to alter a book (without the Author's consent) an then sell it. (And this is a right that the Author can't even give away)

Re:It's not about the patent, it's about the lying

[TheRaven64]

It would be a problem in much of the EU. The USA does not have a concept of Moral Rights at a federal level (although some states do; California has quite a lot, New York does but only for certain kinds of art). In many European countries, especially those influenced by French law, the author's name may not be associated with derived works without their consent. This includes even minor edits. Moral rights can not be assigned in much of Europe and so Amazon would need permission from every author, not just every publisher, for each modified form.

Re:It's not about the patent, it's about the lying

[instance]

Quick, file a patent on "System and method for correcting multiple variants in content" before someone else does.

Re:It's not about the patent, it's about the lying

[thejynxed]

I would argue that we've had a decrease in creative works (excluding the gaming industry) as a whole since the copyright increases were established. The quality of said works has also seemingly diminished in direct proportion to the length and breadth of copyright applied to the works.

It's extremely easy to see this in the visual arts and music industries. Outside of stuff from people like Warhol and Basquiat, for instance, there has been very little to compare to artists such as Rockwell, Monet, etc from a quality standpoint. I would dare say where are all of the good artists hiding at? Are there even any left?

I don't think I delve far into the music industry, as it's been obvious for quite some time the entire thing has been swirling down the toilet bowl. Even the quality of live shows has been degrading over the last few years, with entities such as Ticketmaster and LiveNation turning what should be a monetary and creative incentive directly to and for the artists and fans into a wasteland of corporate stupidity.

Re:It's not about the patent, it's about the lying

[Anonymous Coward]

Yup, I'm (ahem) fairly knowledgeable about this field, and this is a really dumb (and elementary) scheme. They shouldn't be granted a patent on it, as it's a textbook example of what not to do in watermarking.

Why? Because it's noticeable, and they've just told everyone about it. Given that this is a field overwhelmingly reliant on obscurity, that was really not a very smart thing for them to do.

The goal of steganography - and watermarking is a form of steganography - is to remain undetected to an attacker. The most powerful thing a watermark can be is invisible, because it's only got an advantage when the attacker doesn't know it's there. When an attacker knows that a watermark is present, they can use an attack such as a collusion attack (which you described, and which is one of the most powerful classes of steganalysis/antisteganography), although other powerful attacks also exist.

While some schemes attempt to be resistant to this, nothing (even a mark tree) is actually immune, especially if you know how the marking scheme works. It's not even true that the stronger and more noticeable you make a stegomark, the harder it is to remove. Sometimes very strong ones are easy to remove (although weak ones are never hard). And the outcome from a successful colluding removal will be higher-quality and more compressible than all the watermarked copies, because you've removed the (deliberate) noise.

Unfortunately to the would-be watermarker, as a matter of theory, optimal lossless compression provides a perfect stegomark detection, and optimal lossy compression provides a perfect stegomark removal. Most of the compression and filtering techniques out there make absolute mincemeat out of the entire watermarking field, providing for very powerful detection and removal attacks, and they can even improve the quality on the way (in this field, for example, instead of a scanned PDF containing the encoded noise of a watermark, even scattered in written form - which will reduce the compressibility of the text as introducing a stegomark raises the entropy - the attacker could collude amongst 10 or 15 different copies and reconstruct the source text, producing a smaller, completely clean PDF with no unnecessary scanned images or noise; a file of identical or superior perceived quality in a smaller form is generally more desirable).

A more subtle technique of tweaking the typesetting - inter-word tracking, perhaps, or minor changes in line spacing - would be more effective because it is more likely to go unnoticed, but once the attacker knows a watermark may be present, the game is up.

And that's what makes this scheme really dumb; that they've told everyone about it now, so now people know to look for it. Besides, I can't imagine the authors being particularly happy. Their publishers may have rights to edit for copy-setting or localisation purposes, but the authors may retain rights of approval in some cases, and I'm pretty sure that right would never extend to the retailers...

Re:It's not about the patent, it's about the lying

[shentino]

I still prefer accusing them 3 times of copyright infringement, it's a lot easier that way...

Re:It's not about the patent, it's about the lying

[Alerius]

Yes, because file sharing destroyed the music business. There are now no original artists making music any more. I lament that sad day. it reminds me of the olden days, when we still had movies before the VHS machine destroyed that venerable business model. It went the same way as radio after the cassette tape.

Legit customers seem to still put a few dollars into these things, even when free is available.

Go stand over there with the RIAA, MPAA and Chicken Little and hang your head in shame. The sky is not falling.

Re:It's not about the patent, it's about the lying

[lwsimon]

Even the idea that it is possible to do so is enough to get out of prosecution in the US for a criminal violation.

Unfortunately, the standard for a civil suit is lower - "preponderance of evidence". You really could ruin some schmuck's life this way.

Re:It's not about the patent, it's about the lying

[cpt kangarooski]

I would argue that we've had a decrease in creative works (excluding the gaming industry) as a whole since the copyright increases were established. The quality of said works has also seemingly diminished in direct proportion to the length and breadth of copyright applied to the works.

Well, copyright only looks at quantity, not quality. Quality is, after all, subjective, and it's the last thing that we want the government involved in, at least with regard to granting copyrights.

As for quantity, I think that while there has been an uptick, it is probably not all that attributable to copyright. Technological improvements are more significant, as are some other factors. I don't think there's been a decline.

I would dare say where are all of the good artists hiding at?

Oh, that's nothing to do with copyright. The trend in the visual arts for some time now has been to produce crap -- unattractive, devoid of significant (or any) meaning, and intended only to be enjoyed by people in the art world, rather than the general public. Architecture has had the same thing going on since the 1930's or so.

Moral rights

[davidwr]

Canada and some other countries have "moral rights" which belong to the author.

Changing words without his permission could violate these rights.

In some countries these rights are inalienable and non-assignable. This means the author can't be ordered to waive them by the publisher or other copyright-holder.

Re:Moral rights

[Jared555]

They can't be forced to but can someone say 'we will pay you an extra million to give up this right' is that legal?

Re:Moral rights

[mrbester]

Even if that was agreed the rights are still non-assignable, so the author can cheerfully take the money and then sue anyway. It can't be considered breaking a contract as the contract was unenforceable in the first place.

Re:Moral rights

[sorak]

In the US, you should really try mixing the words "moral" "business" and "law"...I think the resulting explosion would be quite interesting.

Re:Moral rights

[dkf]

In the US, you should really try mixing the words "moral" "business" and "law"...I think the resulting explosion would be quite interesting.

The US simply doesn't recognize moral rights (which have a fairly good definition in the Berne Convention, IIRC) when it comes to copyright law. However, the US is also unusual in this respect.

Never mind...

[NotQuiteReal]

I saw another post that already has the wikipedia on exactly this, and, as a bonus, includes:

A Fred Saberhagen Berserker science fiction short story, "The Annihilation of Angkor Apeiron," has a Berserker directed to a star system by an encyclopedia salesman. The salesman is put on trial for treason, but reveals that the encyclopedia article for the star system, with population figures, resources, etc., was a fictitious entry included in the encyclopedia to detect plagiarism; thus the Berserker actually ended up in an empty star system where it ran out of fuel and ceased to be a threat to humanity.

The Authorized Amazon Version of The Bible

[Tynin]

And people complained about the King James version being altered. I can just picture it, 20 years from now, a group of tomorrows theologians are busy studying the Authorized Amazon Version of the Bible trying to deduce the 'real' meaning of the text/God.

Re:The Authorized Amazon Version of The Bible

[mbone]

Let's hope that tomorrow's theologians actually know how to read Greek, Aramaic and Hebrew, and so are not dependent on translations, authorized or otherwise.

Re:The Authorized Amazon Version of The Bible

[Fished]

An enormous amount of effort, and I do mean enormous, has gone into producing a critical text that is probably very, very close to the original. When the Dead Sea scrolls were discovered back in the 50's, with a copy of Isaiah several hundred years older than any that was had previously, there were virtually no differences between it and the Masoretic Text (which is the standard, scholarly text of the Hebrew Bible.) With respect to the New Testament, they literally compare thousands of manuscripts, dating them, slicing and dicing them to determine what the earliest text was.

And, let it be said, there are no major doctrinal issues that hang on a textual question. Translational questions, sure (e.g. should be translated "faith in Jesus Christ" or "faith of Jesus Christ" [Romans 3.22]) but not textual questions.

Re:The Authorized Amazon Version of The Bible

[sheph]

Everyone knows the NIV is the only true Word of God, and all other works proclaiming the nature of God are a product of the devil and must be destroyed. (yes I'm kidding, just in case you missed it)

Re:Don't shop amazon if you like artistic integrit

[godrik]

I was about to say it. If they do that, I'll never buy a book at their e-shop. I don't care about watermarking that does not change the visual quality of a rendering (despite each video technology improvement shows the imperfectness of previous records ), but changing the content is not acceptable. why not change the color of the hair of actors in movie while we are at it ?

Encyclopedias

[JesseBHolmes]

It's a bad practice, but encyclopedias and dictionaries have been doing for years. See Lillian Virginia Mountweasel, the fountain designer/photographer who overcame non-existence to be featured in the 1975 New Columbia Encyclopedia as an anti-piracy measure.

If I was an author . . .

[Tanman]

If I was an author who had slaved a year over a book, and anyone but my editor (with my approval on each change) altered my precious words and distributed it as my work, I'd sue the pants off of them. It'd be like if someone was selling prints of my painting and changing a brush stroke. You just don't do that. Words are the author's paint.

Re:If I was an author . . .

[cdrguru]

As an author, you probably get to do that when you reach the revenue generation level of Stephen King or Issac Asimov. Until you have 10+ books to your name the publisher's editing team is going to do whatever the heck they want to your book and keep the copyright to themselves. It is in the contract you grudgingly agreed to because you wanted to be a published author.

After you sign such contracts for the first ten books, you might just be able to negotiate that your words are inviolate and you get to keep the copyright.

Re:If I was an author . . .

[Jared555]

Eventually it is going to be much more common to self publish for these reasons. Yes it happens a ton already but you aren't going to get your book into all the chain stores by yourself. If you can self publish on a major ebook (or on demand print) store with the option to opt out of this service, keep all rights to terminate distribution, etc. then more people will start doing it.

Re:Don't shop amazon if you like artistic integrit

[mr_zorg]

I can't believe they would actually try to apply this to others' works without their consent. This seems more likely to be used *by* authors. Or at least some of them in certain situations. Could definitely be useful for corporate memos, etc. to find leaks. :-)

Thank Goodness!

[Seraphim_72]

Now Tom Sawyer wont have nigger in it. Unless you want it too. And Lolita will be 'just' 18. So you can watch it legally. Lets all change words to be more acceptable to the audience right? I mean 'What fools these Mortals be' is the same as "ppl r dumb' right?

I thought of it first

[lseltzer]

I was doing this with Cliff Notes 35 years ago

not all that is written is true.

[kras]

be it in stone or the internet. good luck in finding out the liar and the truthsayer.

Cartographers already do this...

[pongo000]

...that is, introduce deliberate errors into their maps to detect copyright violations. Here's an example of an island that was simply "dropped" in the middle of a lake.

eBooks? Thanks, but I'll stick to paper.

[Adrian Lopez]

This is yet one more reason not to get a Kindle or buy any eBooks from Amazon.

Re:Quick! Someone say it's only defensive!

[mysidia]

Two words: One click.

Let me see if I got this right...

[Aldenissin]

You want to prevent piracy by altering parts to later attempt to prove it was pirated? Ok, do it. Then the pirates will just do the same and systematically substitute synonyms themselves in order to delude this technique. Sure it may not be perfect, but this sounds almost like a challenge by the industry. And you know what happens when you tempt hackers with an apple. As the saying goes, "Security through obscurity has never worked, and it never will." But I bet you can still market it to the drones who think they are paying for adequate technology. And if it reduces piracy out of sheer fear, and just gets people scared, then it is akin to "bad" publicity. Although it may be negative, it could still be considered successful.

Like the Dialectizer or the lolcat translator?

[Roblimo]

Hmm? Does this mean Amazon has re-invented and patented The Dialectizer? -- http://www.rinkworks.com/dialect/

Or the lolcat translator? - http://speaklolcat.com/

"SPEEK SOFTLY AN CARRY HOOJ STICK" -- Theodore Catavelt

"Speek sufftly und cerry a beeg steeck" -- Theobork Borkevelt

Not a Wise Practice

[Frightened_Turtle]

First, there is already pre-existing examples of this practice. Indeed, Tom Clancy described this very technique in one of his novels and called it, "The Smoking Word Processor."

Second, as an author, I go through quite an effort to ensure that the spelling and grammar are correct throughout any work that I created. To have Amazon completely throw away my efforts and ruin my work would really anger me. This might encourage me to inhibit Amazon from selling any of my work.

Re:Not a Wise Practice

[arthurpaliden]

So someone distributing the altered work would not be in violation of your copyright because you have not copyrighted the altered work.

Re:Not a Wise Practice

[Mhrmnhrm]

Errmm... I could have sworn that Clancy called it "Canary," not "The Smoking Word Processor." Either way, it's 20-year old prior art.

Re:Not a Wise Practice

[slicerwizard]

"First, there is (sic) already pre-existing examples of this practice."

"Second, as an author, I go through quite an effort to ensure that the spelling and grammar are correct throughout any work that I created (sic)."

Hm. Convince me.

Re:Not a Wise Practice

[vidnet]

Second, as an author, I go through quite an effort to make sure that the spelling and grammar are correct throughout any work that I created. To have Amazon completely throw away my efforts and ruin my work would really anger me. This might encourage me to inhibit Amazon from selling any of my work.

Modifying spelling and grammar was just a secondary modification for patent blanketing. The main point was to replace words with their synonyms. You or your editor could go in and find 30 places where a word can be substituted for another, and you'd be able to generate a billion combinations. Chances are you even read the part I quoted without even noticing my modification.

Re:Not a Wise Practice

[Frightened_Turtle]

I prefer to "ensure" rather than "make sure".

There are times when I want a character to speak in a certain manner. Or have a narrative to sound a certain way. To replace the words I chose with their synonyms could effectively change what I wrote.

Byron's words, "She walks in beauty, like the night," do not sound as good when it comes out, "She strolls in beauty, like the night."

The chances that their system has read one of my novels, understood the nuances between the characters and their manners of expressing themselves and comprehended the scenery and feeling that I was trying to convey before inserting 'a combination of words' is highly unlikely. Unless, the programmers at Amazon have effectively created an artificial intelligence on par with human sentience.

To reduce writing to a mere combination of words would be on par to saying that a trained chimpanzee can manage a video network by pressing certain buttons when they light up. Or that any high school kid could create a corporate web site using a drag-and-drop interface. Can you write a program by copying and pasting code from other people's programs?

Try reading any of the Harry Potter books and if Harry suddenly raised his baton instead of a wand, you would notice. Of course, it could be worse!

Re:Not a Wise Practice

[Frightened_Turtle]

Yup. Even I miss on occasion. That's why I have people proof read for me. :)

Re:Not a Wise Practice

[Frightened_Turtle]

Yes, Clancy did call it Canary, in the novel it was put into action by the characters as part of the plot. But when he first described the concept in an earlier novel, the character was equating it to "... a smoking gun. Think of it as a smoking word processor..."

It was already done decades ago with log tables.

[hendrikboom]

Even tables of logarithms have been doing this ages ago. The preface in one table informs us that a few places where the suppressed digit is a 5, the table entry is rounded in the wrong direction. This slightly increases the potential error from 1/2 of the least significant digit to something like 0.55 of the least significant digit.

So any serious pirate group

[Maxo-Texas]

Acquires two copies of the work in question. Merges the differences- compares those lists and generates a copy that fingers someone else or no one.

Re:So any serious pirate group

[agnosticnixie]

Yay, I knew these philology classes would pay off!

Re:So any serious pirate group

[mbone]

Wouldn't work, at least if they do it right, and the text is long enough. This is a form of steganography, and what you are proposing is actually hard to do. Imagine Amazon makes, say, 1024 changes in a text. Each change can be regarded as a binary bit (being either present or absent), so each text copy has a unique 128 byte number hidden in it. If Amazon gives out 4 million copies, that means they only need about 24 bits to uniquely identify any individual. If I get my hands on 5 copies, and can correct each error where the text differs (which I frankly doubt could be always done) then I am doing an AND on all of my 1024 bit numbers, leaving 32 bits intact, and Amazon could determine where all 5 copies came from.

If Amazon has information about who is likely to collude, they can make this even tougher by assigning codes based on this information. (For example, the first bit could be one for all Slashdot users, and zero otherwise, so no matter how many of us collude, Amazon could still say if the pirates came from Slashdot.)

Re:So any serious pirate group

[Maxo-Texas]

That's really cool and informative.

So then it depends on there never being a mass market print edition but would be fairly secure as long as it remained electronic.

Hmm and they would have to have a solid credit and never allow gift cards. If they allowed a cash sale, then the best they could do would be to invalidate the kindle.

Does it matter if the 1024 changes are in the same location (as a true bit-- either this word is "red" or "scarlet" or just two recorded changes "5th page, this word was changed to black from sable" vs "5th page, last sentence repeats "the the" on the last page."

Of course, spelling and grammar errors would be easy to clean automatically.

hmmm

xxxxxxx
xxxxxxx
xxxxxxx
xxxxxxx
xxxxxxx

vs

Xxxxxxx
xxxxxxx
xxxxxxx
xxxxxxx
xxxxxxx

Re:So any serious pirate group

[Maxo-Texas]

However,
I think I see a problem here.

What I can get out of both books, is the words where everything is the same, and the words which differ.

Now I can modify the words that differ to a new value and I can correct any spelling or grammar errors.

So
XYXXXXXDXXXX22XXXXXXYXXXXXX
XYXZXXXXXXXXXXXX11XXYXXXXEX
becomes
XXX?XXX?XXXX2XX1XXYXXXXXX where Y is a code but I can't tell because its the same in both books. 1 & 2 are trivial duplicate word sets ( "the the" on a page boundary for example). E is a spelling error easily fixed.

It seems that, at best, your 24 bits would identify a set of books, but not an individual book. Because- ANY differences are wiped out. So the only information that survives are bit values that were the same in both books.

What am I not seeing?

Re:So any serious pirate group

[Maxo-Texas]

Dang it..

XYXXXXXDXXXX22XXXXXXYXXXXXX
XYXZXXXXXXXXXXXX11XXYXXXXEX
becomes
XYX?XXX?XXXX2XX1XXYXXXXXX : where Y is a code but I can't tell because its the same in both books. 1 & 2 are trivial duplicate word sets ( "the the" on a page boundary for example). E is a spelling error easily fixed.

After manual patching becomes

XYXAXXXWXXXX2XX1XXYXXXXXX where A and W are manually chosen synonyms.

Even worse, with three sets
XEXXX
XXXYX
XXXYX
-----

I can now create a version that has XEXYX which may pin some unknown person as the source.
Perhaps in regular stegonagraphy (sp), you don't have multiple copies of the same picture with different encoded data to compare.

Imagine the book as a picture of a rose.
Then make a thousand copies and watermark each copy uniquely.

Also, would the bit caught matter?

For example,
1001 - the first bit determines if it is in the first set of 8 or the last set of 8. The last bit determines if it is in the even or odd set of books. Guess no difference in the magnitude on position unless it isn't a true binary set (say there were 1225 books printed, then there would be one "binary" set divided into 1024 books and another into 201 books).

However, if you change the key to ?00?, the best you can do is know the book was in a given 25% of the total books distributed.

You could probably also force more unique key values by ordering over different days, in different parts of the country or in different countries. Of course Amazon could randomize the sequence the changes would be distributed in to combat that.

ah well. . past bed time.

Re:So any serious pirate group

[selven]

So get your hands on 6 copies. Or 20 copies. Or set up your pirate distribution system so that it keeps adding books to the pot and end up with over 1000 copies.

Re:So any serious pirate group

[selven]

The scheme he's talking about is just a list of addresses (ie. page 5, line 14, characters 40-46; page 18, line 22, characters 28-30, etc).

Re:So any serious pirate group

[Maxo-Texas]

I don't know... just finished brushing up on Steganography and it really seems centered around sending a single message in another message (carrier) which doesn't have a lot of other copies to compare to.

This is about sending many unique messages inside of a common carrier. It seems like anything that could uniquely identify a given recipient would be easy to filter out by comparison of multiple sources. Perhaps I'm just being fuzzy and missing it. But for example: Picture of apple, Picture of apple with message added would fall easily. What seems to make steganography work normally is the fact there is no base image to compare to.

Perhaps this is more targeted at casual or idiot sharers?

BRILLIANT!

[chill]

We can set the copyright lawyers, representing the authors and publishers, against the patent lawyers representing Amazon. With any luck, they'll sue each other into the poor house and leave the rest of us alone!

Alternatively, we could establish a special court that handles these copyright vs patent cases. When all the lawyers arrive, wall the area up, cut the bridges and toss in a few spiked baseball bats to let 'em fight it out with. Maybe in New York...

Re:BRILLIANT!

[chord.wav]

It would surely become a reality show... And if I get the chance to see a lawyer being beated up to death I'd surely watch it.

Re:From an author's perspective

[Grimbleton]

So.... what have you written?

How did this get through

[mbone]

Mapmakers have been adding fictitious towns for many years (as many have commented).

People who sell lists have been doing this for many years. (Who's Who, for example, adds a few fictitious people for this purpose, and I believe so do the Yellow Pages.)

People trying to catch spies have been doing this for many years. (I first heard about this during the Thatcher years in the UK, and it wasn't new then.)

So, how, exactly is this new and non-obvious ?

Re:How did this get through

[stephanruby]

Mapmakers have been adding fictitious towns for many years (as many have commented).

Damn, I've always wanted to become the mayor/sheriff/owner of my own town.

So they patented paraphrasing?

[mark-t]

Wow.... just, wow.

Bezos principal

[tuxgeek]

The new Bezos contest: Who can be more evil

Water goats...

[marciot]

Great. I'm looking forward to a whole new crop of engineering textbooks with references to "water goats" instead of "hydraulic rams"

Re:Don't shop amazon if you like artistic integrit

[stine2469]

Two words: Ted Turner.

The mousetrap

[westlake]

This idea has been around forever - and it works.

The plagerist - the infringer - is almost by defintion a lazy son of a bitch. Reviewing text line-by-line. The movie frame-by-frame. That's hard.

Uhm copyright violation through derivative work

[RichMan]

Unless they have specific permission from the owner of the copyright work for any such modification. Any operation such as this would be an unauthorized derivative work and be in violation of the original copyright. The variations would be derivative works, not works in their own rights. Their creation would have to be authorized by the owner of the original copyright material.

http://en.wikipedia.org/wiki/Derivative_work

Subject to sections 107 through 122, the owner of copyright under this title has the exclusive rights to do and to authorize any of the following: (1) to reproduce the copyrighted work in copies...; (2) to prepare derivative works based upon the copyrighted work; (3) to distribute copies...of the copyrighted work to the public by sale or other transfer of ownership, or by rental, lease, or lending....

Re:Uhm copyright violation through derivative work

[MyFirstNameIsPaul]

So those who want to self-publish through Amazon better read that agreement very carefully...

Re:Uhm copyright violation through derivative work

[cdrguru]

Yes, but this is likely happening on the Internet, a pretty much law-free, consequences free zone. Nobody on the Internet pays much attention to copyright, so it is only realistic that corporations are going to start taking advantage of this.

If Russian hackers can steal your bank account and nobody can do much about it, expect to see Sony stealing your music compositions and selling them on the Internet soon. If college kids can download movies, expect Netflix to start downloading them and offering them for rental. Should it surprise anyone that Amazon might be doing something with books that might be questionable?

The logical progression

[Impeesa]

If this becomes widespread, here's how it'll go: first, pirate groups will only have to pay for/obtain a couple extra copies, and come up with an automated reconstruction system that will compare the copies and perform error correction. Then the publishers will start obfuscating things more and more, and the pirate groups will develop more and more advanced algorithms. Eventually, the publishers will be publishing near-100% noise, with their heads too far up their asses to realize it, the only people buying copies will be the dedicated pirate groups, who will afford it by charging for their services, and before you know it, "content miners" will just be another step in the chain. The establishment is just last generation's rebels, am I right?

Re:The logical progression

[Anonymous Coward]

You underestimate how evil a watermarking algorithm can be. Rather raising the number of words changed, amazon can simply make sure one set of an official copy's edits are unique, but another set overlaps exactly with group A of other accounts, another set overlaps with group B, another overlaps with group C... such that a naive copier will still be caught, and collaborators will never be able to be completely certain they removed every watermark.

Then amazon builds sets of potential pirates for each book they find in the wild that was only partly dewatermarked. (In other words, the set of accounts who shared only those remaining word edits in common). This alone will catch a few, but it's not the true strength of the approach. The true strengths are that 1) pirates have to spend more money on copies, discouraging copying 2) the MORE pirates collaborate on a book, the SMALLER the set of potential pirates is if any watermarks were missed.

Then amazon compares the potential pirate sets to each other, looking for associations between groups of users. Pirates would be huge outliers in the results. The number of potential-pirate lists your account is in is important, as is the ratio of total books you purchased vs total potential-pirate lists you're in. If three people each had 100 purchases in common AND were in the same potential-pirate lists for all of them, then they've been busted.

The greatest benefit of this approach is that while amazon can't catch *everyone*, they'd catch prolific ones, and since the punishment goes up with the number of violations, they end up with some high profile cases that do a better job of discouraging potential pirates.

The downside, of course, is that we normal people get mauled books whether we buy them digitally or download dewatermarked copies.

Re:The logical progression

[KlaasVaak]

You underestimate how evil a watermarking algorithm can be. Rather raising the number of words changed, amazon can simply make sure one set of an official copy's edits are unique, but another set overlaps exactly with group A of other accounts, another set overlaps with group B, another overlaps with group C... such that a naive copier will still be caught, and collaborators will never be able to be completely certain they removed every watermark.

This only works if Amazon is the only content provider though. I dont know any author that sell their books exclusively via Amazon so the only thing you'd have to do is to get a copy from a different publisher.

diff

[KamuZ]

Someone buys 2 books (same), should be different right?
So make diff on them, mix some words from one and the other randomly.

Then you have a new copy.
People always find a way. Just see in YouTube all the people using cams to show you a video so they don't match the video or audio recognition.

Rollovers prior art

[Tablizer]

They essentially patented rollovers, the kind you get with an HTML "title" attribute used with Div, Span, etc. If not, then just use "title" as a work-around. (Not all browsers support it.)

Re:Copywrite violation?

[Macgrrl]

And would the resulting document be considered a derivative work?

What are they trying to 'fix' here?

[SoonerSkeene]

Everyone here is focused on how illegal this patent might be in practice (changing the author's words). But my question is more fundamental... What are they trying to prevent? Is it illegal for me to loan to a friend a copy of a book I purchased?

For those who can't understand the article ...

[ProfM]

From the JiveSpeak Translator

"To 'esist o' not t'exist, dig dis: dat be de query. Slap mah fro! Dat's whut de famous Hamlet soliloquy might look likes if subjected t'Amazon's newly-patented System and Medod fo' Markin' Content, which calls fo' ' honky codematically substitutin' synonyms into distributed text content,' includin' 'scribblin's, sho't sto'ies, product reviews, scribblin' o' movie reviews, news articles, edito'ial articles, technical sheets, scholastic sheets, and so's on' in an effo't t'uniquely identify customers who redistribute material. In its descripshun uh de 'invenshun,' Amazon also touts de use uh 'alternative misspellin's fo' selected wo'ds' as some way t'provide 'evidence uh copyright infrin'ement in some legal acshun.' Afta' all, anti-piracy measho' nuffs should trump kids' ability t'spell co'rectly, shouldn't dey?"

Oops ... just killed the patent.

Re:Copywrite violation?

[supernova_hq]

It's a copyright, not a changeright.

Is it actually legal? It's definitely wrong.

[meerling]

Would the author consider this as some form of plagiarization?
After all, the author has probably only given them permission to distribute his work, not to distribute numerous altered versions.

For that matter, using synonyms can actually change the feel and meaning of a sentence when viewed in context of the whole.
And for documents relying on factual materials, quotes, and many sciences, swapping out words for synonyms will completely destroy the statements.
Just imagine this for your research, "Fermilabs has discovered a new capture that has a reel -3/4".

That just doesn't work, it's completely wrong, and should never be done, but is a real possibility with an inane patent like this one.
If Amazon actually starts doing that, they can kiss goodbye to sales from anyone who desires unmangled books.

Re:Is it actually legal? It's definitely wrong.

[Joce640k]

Wasn't there a big fuss made when film studios started altering swear-words in movies?

If I was a book author and people were using computer-generated synonyms in my books I'd be even angrier...it's far more destructive than changing 'damn' to 'fuck' at the end of Gone With The Wind.

idiots

[Tom]

First of all, this is, of course, a copyright infringement itself because Amazon does not hold copyright on the books it sells, the authors and/or distributors do.

Two, it's the dumbest thing I've heard in a long time. One, it will not stop piracy in the least. Contrary to the Canary Trap already pointed out in comments, e-book pirates will very much have access to several versions of the book. A simple diff will fish out the "key" words. Then you can mix them so that at least they don't point to any of your actual sources.

Three, it's another example of reducing the quality of your product in order to max your profits. Also known as "punishing the honest customer". That's a really good idea... not!

This patent guarantee's I'll never buy one

[kawabago]

Now you can be tracked via your book! What a great idea if you're Big Brother! If you're a reader this should put a last nail in the coffin of electronic books. Anyone who buys an ebook now is an idiot.

hair splitting

[jipn4]

But no, they did not patent *doing* this, they patented the *way* that they do this.

You're incorrectly assuming that a common shorthand for talking about those kinds of patents implies ignorance of the patent system.

To spell it out for you: the "way" they patented this is an obvious engineering solution to the actual problem they are trying to solve. If you gave the problem of "alter the text so that each customer gets a unique copy" to a CS undergraduate, this is the kind of engineering solution they'd come up with.

(Actually, the first engineering solution they'd come up with is to alter the whitespace.)

Re:hair splitting

[HungryHobo]

Zero width spaces are great for that by the way.
(CS undergrad who's had to deal with the problem)

Re:hair splitting

[jipn4]

If we're talking Unicode, the possibilities are nearly endless since there are so many denormalized forms of Unicode.

I think Amazon is talking about something that survives displaying. In that case, either changing the whitespace or moving the characters around works.

If it needs to survive OCR, then misspelling, synonyms, or changes to punctuation are the obvious choices.

Amazon, I dare you...

[ctrl-alt-canc]

...to apply your patent to this work.

Prior art?

[peterwayner]

I haven't bothered to read the patent application, but there's a brief description of this in my books Disappearing Cryptography and Digital Copyright Protection . In addition, Mikhail Atallah's group at Purdue has explored many similar ideas:

  http://www.cs.purdue.edu/people/faculty/mja/

Suggestions....

[Loki_666]

1) Take the original, convert to PDF, OCR back to text. Voila... enough errors to make it untraceable.

2) Hack into the head of Amazon's computer and download HIS copy of a book. Now release via torrent and wait for the witchhunt to start. He is obviously guilty because it is HIS copy that was pirated.

Simple solution!

[KingSkippus]

It seems to me that there's a pretty easy way to defeat this. Use the technology against itself.

If you ever want to distribute something, make your own minor spelling variations and substitute your own synonyms into the original, thus further altering the altered work. If someone sues you, just point out the fact that their copy "proving" you're guilty doesn't even match the copy of the work that was distributed.

You could use this idea for just about anything that is digitally watermarked. Don't want that MP3 traced? Introduce your own small, imperceptible variations into the waveform. Don't want your printer tracing you through microdots on your hardcopies? Write a driver that adds its own microdots, and lots of 'em. And so on...

Re:Simple solution!

[pvera]

Your microdot idea is downright evil, I love it.

Derivative work...

[Kindaian]

By changing the work, as it's in an intended form (as they make sure the change happens on purpose), so they are really creating a derivative work.

That is not a typo... it's a intended change on purpose.

Did they have the author authorization for doing such and at the same time imply that the author doesn't know how to write???

I hardly believe that...

This is going too far...

[Genda]

Never again will I ever bye a book from 'Large-single-breasted-female-warrior.com'.

Re:A Scanner Dumbly

[Genda]

Forgive me, but "A long stemmed flower by any other moniker would scent as agreeably" would be fair grounds for having Shakespeare's ghost kick your ass for an indeterminate length of time. There are words who's subtle and beautiful composition should not be screwed with, particularly in the wanton attempt to prevent people from sharing and appreciating those words (simply because you think it may impact your business model.) You want to finger print software texts, be my guest. You want to screw with my E. E. Cummings, then we're gonna have some problems.

Prior Art - Violation of Copyright

[EmagGeek]

Okay, so aside from the copyright violation this may very well be, map makers have been doing this for centuries. They make subtle yet innocuous errors in maps to make sure that other map companies don't steal their work for their own. How is this any different or innovative? Oh right, it's not.

Substitute this, Amazon!

[CuteSteveJobs]

I always look at my Bezos before I flush.

an idea

[Mantorp]

fix all the instances of 'should of' and 'could of' in the Great Gatsby

Re:Don't shop amazon if you like artistic integrit

[mathx314]

Actually, I've seen the idea of doing something like this with movies before here on Slashdot, and thought it was a clever way to try to pinpoint the location of leaks. Say the main character in a movie drinks a soda at some point. Have the can that he actually lifts up to his mouth be blue, and then edit it to be a Coke or Pepsi in two different versions of the final reel. Repeat this with, say, 16 different unimportant changes and you've managed to encode your movie with a 16-bit pattern that unambiguously identifies each one of the original reels. Now when your movie leaks before release date, you go and download the copy, watch it, and figure out where the leak came from. I think that just so long as nothing significant was changed nobody would mind too much.

Why not just modify whitespace?

[jeffb (2.718)]

Develop a robust, redundant code around substituting two spaces for one after a period/full stop, or adding spaces at the end of a paragraph. Or occasionally substitute em- and en-spaces. It accomplishes the goal, and it doesn't vandalize the source.

I certainly hope there's more than enough backlash to kill the idea of changing words. Otherwise, after ten or twenty more years of natural-language understanding research, we're going to see them trying to do the same thing with plots, or character development.

Re:Why not just modify whitespace?

[DigitalSorceress]

Yep, that's what I was thinking too.

Of course, my concern would be that it would then be kind of easy to strip back out... all you'd have to do would be to run

s/  / /g

until it came beck with no matches.

If you think they may be substituting other white-space characters, come up with a standard for mapping that out too.

What a terrible idea

[sheph]

I wonder if they ever stopped to consider that writers choose what words they use to convey a concept rather carefully, and that they might not have an overwhelming appreciation for a scheme such as this. Is it really so important to catch someone redistributing that you have to stomp all over the integrity of the work itself? If it diminishes the value of the work then doesn't that work become worth less? These are questions that they ought to be asking themselves before venturing into such an obsene effort. I certainly have no interest in reading a book with this scheme applied to it.

Will this change the meaning?

[Name Anonymous]

Many words have multiple meanings. Will Amazon pick the synonym that has a compatible meaning? Or will they change the work totally?

Also what if a word they pick is also the name of an item? This would break the work they modified.

An example works for both: "Grill" do they mean to cook or to interrogate? And what if it is a grill as in the item to cook with?

And as for changing the spelling of words, well what if it becomes another word? Or maybe it gets changed into a name used in the book - this would cause confusion. And what about people learning the wrong spelling for words?

Of course there is also the issue of possibly violating the authors copyright by changing the work in question.

Well this encourages piracy

[Atrox666]

If I pay my own hard earned money I get a fucked up copy of the book I bought?
If the pirates cross reference out those errors then that means the pirates are handed a monopoly on the original work as the artist intended. Hand you opponent a monopoly on your own product. Genius plan.

Amazon Patents Changing Authors' Words

[sirrogerdecoverly]

The Works of William Shakespeare with additional dialog by Amazon

It's also about art

[OhHellWithIt]

This is something that should really tick off writers. An author puts a lot of work into choosing the right words for what he wants to say and wrestles with his editors over word choices until the work is finally in print (or published, by whatever means). Now Amazon is going to take a damned computer and change the words willy-nilly? Please!

Of course, this may someday lead to an ironic situation in which someone purchases a digital copy of a book or something, and one of the most oft-quoted sentences in the work has been changed, so that the reader is cheated of finally reading the sentence in context. I can see it now, the famed "Out, damned spot! out, I say!" from MacBeth, changed to "Out, wretched stain! out, I say!" Or, perhaps worse, the quote with the changed word(s) becomes more widely known.

Re:It's also about art

[omnichad]

Can't believe I had to scroll down so far to see a comment like this. I would hate it, too. For the same reason. I don't even like watching movies on cable because they've been edited for length. Or had poor dubbing on certain replacement words.

Preemptive Patenting?

[MaliciousSmurf]

Not to be a heretic, but, wouldn't you rather a company like Amazon hold the patent than some random patent troll? In the current climate, I could see them patenting some kind of similar (bullshit), charging a nominal fee to license the patent, and everyone's happy. (As compared to some patent troll getting their hands on the patent and suing God, Jesus and the Holy Ghost for infringing on their "triune divinity" patent.) As mentioned by a previous poster, btw, this method has been in use for years. Like, hundreds. Canary Trap. Wiki.

New flash for Amazon

[FreekyGeek]

Amazon also touts the use of 'alternative misspellings for selected words' as a way to provide 'evidence of copyright infringement in a legal action.'

They're claiming to have invented the Canary Trap? The prior art on that is 30+ years old...

Re:Americans say "May the 1st", not "the 1st May"

[PriceIke]

Cinco de Mayo,

Canary Trap

[thickdiick]

The canary trap is easily defeated by rewriting leaks in "your own" words.

Just summarize the classified document in a writing style randomly similar and dissimilar to your own writing style.

Only if...

[eth1]

...I can sue Amazon for damages when my "Hamlet" paper is docked points for inaccurate direct quotations.

as if it would really help

[v1]

if you REALLY wanted to copy and distro it, just buy it with three accounts. Do a simple compare and find the few words changed, and rebuild the original and distribute that instead.

This approach doesn't work for maps for several reasons, but would be ideal for books.

I've already seen this in action where watermarks are placed on digital software downloads. Just a matter of obtaining it from two different sources and comparing the two to find out where the watermark is. Then either remove it or change it to something else. One of these days they'll get smart and start signing it after they watermark it, but that's processor expensive so guessing they don't feel quite that motivated yet.

You can always tell

[mea37]

You can always tell whether someone is trying to preach to the choir, or whether they really want to convince others of their viewpoint. It's too bad so few understand that publically making arguments that serve the former, actually hurts their cause with respect to the latter.

"After all, anti-piracy measures should trump kids' ability to spell correctly, shouldn't they?"

Really? "Won't somebody think of the children?" Kids can't be taught to spell if there are mis-spellings in some material they might see? Get off it. Strained argumetns like these only make you and your entire cause look desperate.

The problem with this technology is, if I buy a book and you change a few words to synonyms, then you haven't sold me what I bought. You've sold me a defective product, and I should be within my rights to sue you for it. (And yes, I think the same applies to mapmakers. If I buy a depiction of an areas geography, I expect every mark in that depiction to reflect a fact about that area. If some marks appear to but don't... you may think it's a "minor" change but it might make a difference to me. Map is defective should mean lawsuit.)

On the other hand, technology like this would be useful for enhancing full-text search capabilities. It's never about the technology, it's how you use it.

Re:A86, anyone?

[drenehtsral]

That's the first thing I thought as well! Those were the days...

Prior Art

[aitikin]

I'm 95% certain that a music publishing company has been doing this for decades. I don't remember which company off hand, but on every page there is one wrong note. It's a very respectable company as I recall.

OBJECTION

[thehostiles]

isn't this where it all starts? with one word here, one there. after a few "generations" of changed books, the originals may lose context and meaning. I'm maintaining a healthy skepticism of this being amazon's plan, to eventually censor books it doesn't agree with and profit from it. Hiding behind anti piracy and protect the children can only last so long... I'm not saying this IS their plan, just that it could easily be possible given their recent actions (suing people for having public domain books (animal farm and 1984 actually) on kindle devices)

HoneyTraps

[omb]

HoneyTraps, have long prior art, Jack Ryan re-invented it in one of Clancys books, but it is attributed to David Niven, then of SIS, not Hollywood c. 1942; but if some idiot at Amazon sells me the wrong book, or one with deliberate misspellings I will enjoy sueing his ass off for 'passing off, or goods not of merchantable quality'.

Why does not a day go by without some corporate jackass trying some stupidity like this, is this all MBA courses in the US teach?

Re:HoneyTraps

[russotto]

HoneyTraps

Canary trap. A honey trap is something different, and I don't see Amazon patenting the idea of using attractive women to lure men into compromising situations. Even the patent office can find the prior art there.

M$ should fear this patent

[ntime60]

Everyone knows M$ has been caught more than once copying functionality of other OSs and suspected many more times. You think they might actually create something now?

Re:Americans say "May the 1st", not "the 1st May"

[lwsimon]

No, we say "Independence Day" :)

"The Fourth of July" is a bastardization of the holiday.

THAT explains the typos!

[psydeshow]

Oh, this is hilarious.

Kindle books are riddled with typos, presumably caused when print editions are scanned to make e-books. (Why don't they get electronic gallies from the publisher? Who knows?)

So either they have been causing them on purpose to track redistribution, or this is a fine example of making patented lemonade from the technological lemons produced by their scanners

I love my kindle, but I hate Amazon more and more each day.

Your idea won't work if ..

[roguegramma]

Your idea won't work if the falsified section of the work always are falsified with 50:50 chance.
On the other, you would be able to put together fake marks easily.
However, a countermeasure to fake marks is to only use a subset of the marks, for example those that match a number produced by an algorithm or have a certain checksum or hash with certain properties.
(I hereby dyspatent that idea by publishing it ;-)

The patent by Amazon at the core is not a new idea, it is just mixing A and B and patenting that feels somewhat silly. There's a guy a round somewhere who works with a matrix method to find such new inventions. I would love to see his website, but I forgot his name ..

Traitor Tracing

[__aayurq3262]

This method is a solution to the problem known in cryptographic circles as "Traitor Tracing." The patent sounds awfully similar to the traitor tracing method used in Blu-ray discs and the old HD-DVD discs. It's a capability of the system licensed by the AACS Licensing Authority for encrypt those discs. Basically, they can substitute one of multiple short chunks of video at multiple places in the movie. After a decrypted movie is released, they can figure out what system was compromised. Interestingly, the traitor system has never been implemented, even though all licensed players must be able to handle it. That's probably because the AACSLA knows what system was compromised - it's one of the software Blu-Ray players.

The software players are all identical. The hardware players can be tracked down to a specific player. Isn't that nice to know.

Finally!

[radicalrendell]

Finally, Amazon can put to use their secret warehouse full of monkeys and typewriters!

They stole my idea!

[GrantRobertson]

Honestly, I did think of this sometime around 2004. I may have posted something about it on Usenet. I will take a look. Others could search too. I have no idea what newsgroup I would have written it in, perhaps a Microsoft one. I have always used the name "Grant S. Robertson" for my Usenet posts. That should narrow things down a bit. Also, I thought of it in connection with publishing and distributing e-books, primarily in .PDF form if that helps.

This is one of the reasons I have started my www.ideationizing.com blog. Simply to post my ideas which I don't expect to be able to work on, and thus provide prior art to stymie patents such as these. I urge everyone to blog and post as many ideas as they can think of. This seems to be the only way we are going to prevent the big corporations from patenting everything under the sun.

Re:They stole my idea!

[GrantRobertson]

I just found the text of my notes that I wrote up when I thought of this. I am posting them here, not because I think this establishes prior art, but because it may help in finding any usenet posts I may have made about it.

Text document serial number encoding scheme

Sunday, January 25, 2004

4:46 PM

1. In several places within the document, identify multiple different possible wordings.
2. • Each location equals a digit
   • Each different possible variation is another character for that digit.
   • • Since there may be more than 2 variations for some locations and each location may have a different number of variations, the serial number is like a byte of data where each bit can be in a different base number system. This vastly increases the total possible different combinations. A 3 bit, base 2 number (binary) gives 8 different combinations. If you change just one of those bits to base 3 then the total possible combinations jumps to 12. for each additional variation in one particular digit it increases the total combinations by the total combinations of all the other digits. This is basic base number system math but not everyone thinks of the possibility of using a different base for each digit.

Re:They stole my idea!

[GrantRobertson]

I searched Usenet for my name and only got 112 entries. I know for a fact that I wrote thousands of messages in Usenet over the years. So it may be possible that I posted this idea and it is not turning up in the search. If anyone else wants to give it a go to defeat this patent be my guest.

And yes, you will find some interesting posts that would be embarrassing if I gave a damn. So don't bother with the cracks, I won't care.

hi

[raynorgmail]

:D

Copyright Violation?

[jvkjvk]

I think this would be copyright violoation on Amazon's part.

They are making unauthorized derivative works. They may (and probably do) have authorization to distribute copies of the original work, but that in no way includes the right to create a derivative work and sell that. Such minimal substitution as they do in no way would constitute a new work, at all.

Of course, I'm not a lawyer, so perhaps this is incorrect, but it seems accurate at first glance to me.

 

Greed

[twoHats]

These greedy bastrds would destroy art and literature for just one more buck...