« A Review of John's Yoo's WAR BY OTHER MEANS: AN INSIDER'S ACCOUNT OF THE WAR ON TERROR | Main | Immigration Law Reform »

November 06, 2006


Feed You can follow this conversation by subscribing to the comment feed for this post.

Doug Lay

>> This is unlikely to occur before the digital material is safe from re-copying.

Digital material is never going to be safe from re-copying. Period.


What would happen to publishing companies? Because authors will write, editors edit, agents promote directly to..Sony Publishing? How does Sony choose what books to sell on E-Books?

Nathan Richardson

Eventually, publishers (including Sony) may be unecessary. Writers can distribute directly to readers for micropayments. Dean Levmore could, for example, post a link to a .pdf (or some DRMed equivalent, though I agree with Doug Lay that DRM is at best temporary protection) here on the blog, and readers could download a copy after making a small payment. The payment need not be tiny if demand is high, but it would be presumably less than for a printed copy since materials and middleman costs are lower.

Someone who knows more than I do about the topic might be able to say something enlightening about how this would change the price of books (currently price controlled in some places), redistribute money between writers, editors, publishers, and readers, and maybe even change the format chosen for written documents (in the same way that song downloads are putting pressure on the album format).

In other words, publishers are in danger just like record companies already are. I think you could make an argument that this is even more true of academic publishers, since the combination of their informed audience, existing community links (blogs, etc.) and high price of academic books makes academic customers much more likely to favor e-books than say, Danielle Steele readers.

Michael Martin

Publishers are not going to disappear. The service they provide has for some time been primarily one of sorting through the flood of unmarketable manuscripts in order to find the few that will sell the big bucks. All that will happen is a division of labor whereby more specialized publishers will come on the scene that did not previously exist. Hoi polloi will still get their pulp fiction served up to them -- but with electronically doped polymers rather than the pigment-doped regular kind.

Tim Lee

Prof. Levmore,

I hope you'll take a few minutes to read security expert Bruce Schneier's essay on digital copy protection:


He explains why digital content will never be "safe for re-copying."

saul levmore

I appreciate the comments on E-Readers and digital content. And I have now read the essay that Tim Lee suggested. I am more skeptical about claims that something (in this case copyrighted digital content) will "never" be safe; new technologies can surprise us. Besides, sometimes a hurdle there and a legal restriction there add up to a fairly safe market. There is a great deal of pirating of movies, for example, but still a remarkably robust market for conventional moviegoing. But, more important, we do not need E-readers to threaten conventional publishing. Books can already be scanned and thus digitalized, so that the threat to publishing does not require the original publishers to move to digital books.

Doug Lay

Sure, new technologies can surprise us, but your original post made it sound like the emergence of effective copy protection would be part of the natural evolution of an e-book market. This blithe view is utterly ridiculous, and the fact that all you come back with is "new technologies can surprise us" just emphasizes the ridiculousness.

As for the "hurdle here and legal restriction there," that's not something we need to evolve toward - the DMCA anti-circumvention provision has been here for years. In case you haven't noticed, it has been a complete failure at preventing Internet piracy, and that is not likely to change. Google for "darknet paper" to educate yourself.

Tim Lee

Thanks for taking the time to read the Schneier article!

There is a great deal of pirating of movies, for example, but still a remarkably robust market for conventional moviegoing.

Isn't that evidence that copy protection is not a precondition to the emergence of a digital market? People can easily get pirated movies from peer-to-peer networks, yet most people still purchase DVDs and movie tickets.

Likewise, it's inevitable that people will put pirated books on peer-to-peer networks. As you point out, that will occur even if the publishers don't allow the creation of digital versions of their books, because people can scan paper books and upload the images to sharing sites. So it's not clear to me how making material "safe from re-copying" is essential to the creation of a market for digital books. Obviously, some rudimentary speed-bumps can be helpful, but I think the content industries are making a big mistake in thinking they can develop technologies that prevent determined infringers from infringing. Rather, their goal should be to make it as easy as possible to obtain and use their products, so that people freely choose to buy their products despite the availability of illicit versions of those products.

Kent McKeever

Re: "I will not be investing in companies that produce paper." I wouldn't be so quick.

I have always found it wonderfully ironic that Lexis/Nexis, one of the most successful database companies in the world, was originally created as a subsidiary of a paper company. Did Mead Paper actually have someone in the early 70's who realized that computer development would result in huge increases in local printing? If so, it is one of the seriously under-acclaimed moments of business foresight in history.


Interesting possibilities for the future of publishing.

Whilst it may be that 'safe from re-copying' continues to be a problem (in the sense of recovering rents), there is great scope for turning a profit in other ways...advertising on the side of a page, links to other books by the author, links to other authors, 20%off a book, etc. I should think that firms will not sit by whilst their profit margins shrink...no more than the first 20 minutes of a movie at the theatre is the intro of the actual film.

Nathan Richardson

Tim Lee's comments on the DRM issue hit the nail on the head. The example of movies is a great illustration of the fact that effective DRM is neither necessary or sufficient for a (non-pirate) market, which is contrary to what Dean Levmore seemed to say in the original article.

Schneier is great on these kinds of issues.


Mr. Lay, I disagree that security is an impossible goal. We already have extensive secure communications on the Internet, and there is no reason to believe that security protocols cannot be applied with equal effectiveness to e-books. Think of it as a simple communications problem: how do we get information to travel from one location to another in a way that is secure from interception? When you think about it this way, security seems an easy problem. The trick is to insure that the receiving end is just as secure as the transmitting end. I think this will be easily accomplished by burning security protocols into silicon. A hacker can always reverse engineer a program running on a known processor, but when the key processing takes place inside a chip, it's REALLY hard to intercept!

Doug Lay


Sure, securing e-book content seems like an easy problem, when you present the problem as something completely different than what it actually is.

If you look at any textbook on encryption, you'll see a basic framework that involves three individuals - a sender, a recipient, and a would-be eavesdropper. The eavesdropper has the ability to intercept the scrambled symbol during transmission, but at no time is the eavesdropper entitled to view the unscrambled message, nor does the eavesdropper have any sort of access to the encryption/decryption key.

With copy protection schemes (or what the marketeers call DRM) the would-be eavesdropper and the recipient are the SAME PERSON. The sender has to allow the user to view the communication AND prevent them from doing something else with it. It is as if Bob needs to send a message to Alice, while preventing Alice from accessing the message.

Basically, the problem of preventing unauthorized copying of digital content is completely different, and much harder, than the basic problem of encrypting communications. Anyone claiming otherwise is either selling snake oil or setting themselves up to buy snake oil.


Mr. Lay, my point is that we can hard-wire encryption into the e-book processor, thereby making the recipient the processor, not the customer. The processor receives and stores encrypted information, and the decryption process takes place on the fly. Why can't such a process work?

Doug Lay


What do you mean by decryption taking place "on the fly"? At some point the unencrypted text is made available to the customer in a persistent manner, otherwise the book is completely useless to the customer. If the text can be read, it can be copied. Even if the channel from the processor to the display can be secured (a big if), AND the key proves impervious to black-box reverse engineering (another big if) the text can be photographed and OCRed, no? And unless the device ONLY supports "protected" content (something that would greatly reduce the device's value), a single user cracking the protection anyhere in the world can lead to unprotected copies available worldwide within hours.

If you still feel unconvinced, I ask you - can you point to any substantial body of digital content anyhere that is not available on the black market due to the robustness of the surrounding content protection?

saul levmore

I don't think these disagreements have much bearing on the world of E-Readers because the piracy problem described, for example, in the previous comment is just as powerful when the pirate must scan a hard-copy book and then send unprotected copies around the world. There may be some data that creators or publishers choose to distribute or not (in digital or other form) because of copying fears, but "regular" books would not seem to fit this bill.


Mr. Lay, you're right that the output from the display can be scanned, but Mr. Levmore's observation that the same thing can be done to paper books is a crucial observation. Piracy presents us with a serious problem only when it is convenient.

By "on the fly" decoding, I mean that the content is stored in encrypted form and only decrypted a page at a time.

You ask whether I know of any digital content that has defied piracy. As a matter of fact, I do -- I wrote it. I do not wish to reveal my identity by describing the software in detail, but about 15 years ago I wrote some consumer software. I was perturbed by the vainglorious claims of hackers that they could crack anything. So I set myself the task of writing an uncrackable piece of code. However, I did not seek to achieve theoretical uncrackability. Instead, I chose to build in a large number of deceptive practices that would befuddle, not block, the hacker. I was completely successful in this -- nobody ever succeeded in cracking the code. I even hired a hacker to spend a hundred hours attempting to crack it. At the end of the hundred hours, he hadn't even cracked the first layer of protection.

Again, the software was not mathematically uncrackable. But it was humanly uncrackable. The techniques I used were a great deal of fun to design -- they all rely on "sleight of hand" in software design; deception; deliberate offering of apparant solutions that are actually traps. I will share one of my tricks, because it was so much fun: when one guard routine detected that its client routine had been altered, it waited until a significant amount of human activity had been carried out -- then reset the client routine to its original form. Har har har.

Doug Lay

The piracy problem only reduces to an "analog hole" problem in a best-case (or worst-case, depending on your viewpoint) scenario involving an uncrackable hardware-based encryption system. Such systems are - thus far - nothing more than marketing hype. Erasmussino can (mis)place his faith in such a system if he likes, but here are a couple more things to consider:

- While embedding encryption keys directly in circuitry makes the system harder to reverse engineer, it also makes the system much harder to update. If the keys should need to be updated in the future, the vendor will need to round up all outstanding copies of the reader and mod-chip them. (Yes, firmware is another story, but the increased flexibility of firmware goes hand-in-hand with increased vulnerability to tampering).

- Increased difficulty or not, hardware-based systems can be reverse-engineered. And every piece of functionality added to the reader above and beyond the ability to view the whole book provides more information for the would-be cracker. Search functionality, bookmarking, highlighting - all can provide clues to the underlying encryption scheme. Will all of that functionality be embedded in hardware? Could be a mighty expensive reader - perhaps only the NSA will be able to afford to buy the product.

I make no claim as to whether publishers will make content available via e-books. I hope they do. My claim is only that belief in a future involving robust digital copy protection is very foolish. Yet otherwise highly intelligent people - especially lawyers - seem to be quite susceptible to this foolish belief. Perhaps it is because the alternative - the technological advancements of the last quarter century rendering centuries of copyright law basically obsolete - it too horrible to contemplate. Recall this quote from the late Saul Bellow:

"A great deal of intelligence can be invested in ignorance when the need for illusion is deep."

Doug Lay


We seem to have crossed paths posting.

I don't place faith in anonymous claims. Name the system or I'll feel free to believe you are making the system up.


Mr. Lay, I agree that a full hardware solution is immalleable, and that is its greatest weakness. Any such system will have to be part of some standard, and that standard must be robust enough to last for its required lifetime -- which in the case of e-books could be decades. I am not optimistic that a security system implemented now would be able to accomodate future requirements gracefully.

On the other hand, I disagree with your belief that hardware solutions are easily reverse-engineered. Those transistors are really teensy-weensy these days, and you need very expensive equipment to resolve them. Tapping into sections of a chip to monitor its internal communications is a multi-million dollar effort. Anybody who can afford the effort cannot afford the legal liabilities. Sure you can tap board traces, but that isn't good enough to figure out what's going on inside.

As to your lack of faith in anonymous claims, that's your right. If you don't want to learn anything from me, I don't feel any need to cram it down your throat. Believe what you want to believe.

Bruce Boyden

Saul, I think there is an important difference in how the copy is made. A scanned book is time-consuming to make, usually lower in quality, and worst of all huge in file size. Those three things, particularly the file size, act as "natural" barriers to rampant file trading in books. (Think of the hassle of making a photocopy of an entire book. Not only is it time-consuming and expensive, but it's larger and probably heavier than the original book.) Not so with e-books, at least if they are unprotected. I think the market will need to rely on either the persistence of "natural" limits (for whatever reason, not many people engage in wide distribution of copies), or technological limits.

I won't revisit the DRM issue yet again here, except to note that the test for success for a content protection scheme is not perfection, but mimicry of whatever "natural" limits that impede the undermining content in successful markets and are absent from the market in question.

Doug Lay


Using scare quotes around the key word in your argument shows a whole lot of confidence and clarity in your thinking.

Copy protection in a digital environment is un-natural (no scare quotes). Copy protection in a globally-networked digital environment is doubly un-natural. Laws that mandate digital copy protection in order to mimic the behavior of legacy markets aim to hamstring the future to protect the past. They are anti-evolutionary, not evolutionary.

Also, regarding your thoughts on scanning, there is a radical new technology called "optical character recognition" that you might want to read up on.

David Sucher

The degree to which people will steal ebooks is not only a function of the ease of transfer but of
• the cost of legitimate purchase and
• the benefits of having a registered copy.

If publisher persist in their foolishness of trying to sell a digital copy of a book for $18, then yes people will blithely steal. But when the price is more where it should be -- say $3-4 -- then that problem will be very much less.

As well, I think most are looking at the book in an old fashioned way as a static finished product. The e-book (at least the non-fiction one) is a product which can be "supported" e.g. with real-time updates and links by the author and/or access to the author for questions/comments etc. It is very much a different product than the paper book and buying it for the "support" can be very attractive.

Bruce Boyden

Doug, I find it odd you have any objection at all to the first paragraph of my comment. You seem to be going out of your way to find something wrong with what I said. For one thing, I drew exactly the same distinction you do between technological limits and "natural" limits. As for the scare quotes, I'm marking the fact that I'm not using that term in a particularly literal or rigorous way -- but this is a comment on blog post, so I really think my obligations to be rigorous are minimal.

Re: OCR -- maybe your experience is different. Mine is that OCR is haphazard, at best, and that you lose a lot of quality in making a digital conversion that way. The scanned copies I see available (e.g., HeinOnline, a resource for law professors) are huge, several megabytes for a fuzzy, fax-like copy of a journal article.

Doug Lay


I think your distinction between "natural" limits and technological limits is murky, scare quotes or not.

Would you classify limitations on file transfers due to lack of adequate bandwidth as a "natural" or a technological limit? I would classify such a limit as both natural and technological - it is a natural effect of the underlying technological environment.

What makes efforts at digital copy protection different isn't that they use technology, but that they use technology to fight against the basic architecture of the underlying technological environment. I call digital copy protection unnatural - that probably wouldn't fly in a court of law, but neither would "un-natural." Better terminology is needed.

Regarding OCR, a couple of points:

- with OCR, the visual "quality" of the scanned page becomes irrelevant, as the text becomes available in Unicode format (or whatever text encoding is used). The important metric is the accuracy rate of the scan. The GPO claims 99% initial pass accuracy on documents they are scanning these days - which isn't all that great. 98% is widely considered a baseline for making OCR more effective than re-keying. I am sure Google is doing considerably better than 99% for their Book Search scanning.

- As for document size, Google Book Search results seem to run about 100K per page, whether in PDF or JPG format. Plenty fast enough to download through your average broadband connection.

The comments to this entry are closed.