[dgatwood]

Originally Posted by Sgt.Stubby

My facts are correct. Encryption does trigger DMCA protections, and there is a prohibition on circumventing it. The fact that other access controls have the same effect was not relevant to my point. You're just muddying the waters here.

Breaking encryption isn't necessarily a DMCA violation. The encryption must be used for protecting a copyrighted work. That's why all those printer vendors that have tried to use the DMCA to sue over someone copying their anti-counterfeiting control chips have had their a**es handed to them in court.

Originally Posted by Sgt.Stubby

Regarding the barcode scanner, I don't recall the players involved. I only recall how loosely the court interprets encryption.

IIRC, that one didn't even go to court; they just sent a bunch of C&Ds and threatened to do so.

Originally Posted by SteveEisenberg

The same is true of the checkout in a bookstore. It only inconveniences the people who pay before walking out with a book.

That's a bit of a non sequitur. Providing people a means to pay is very different from creating a technical means to detect people who don't pay.

DRM is more like those magnetic strips inside products that are detected when you walk out the door. They do work against clueless criminals, but lots of people steal stuff in spite of them, and when cashiers fail to deactivate them, they frequently inconvenience real customers. The big difference is that those strips only inconvenience real customers once (or twice if they end up returning the product), whereas DRM potentially inconveniences customers on an ongoing basis, depending on how the customer intends to use the product.

Originally Posted by SteveEisenberg

Locking doors thousands of times a year is a significant inconvenience for me, but negligibly inconveniences people who are really good at picking locks. What's different about DRM isn't the inconvenience level for those playing by the rules, but the social class of the lock-pickers. Pickers of physical locks tend to have dropped out of high school. But DRM scofflaws fill the halls of UCLA, Wisconsin, and Harvard.

With door locks, you don't risk losing your entire car if something damages your key. And locking your door doesn't prevent you from changing the seat covers or replacing the radio. So the inconvenience level for those playing by the rules does, in fact, differ considerably in the worst case. Car locks are a minor nuisance for legitimate users; DRM limits legitimate users' ability to use what they have paid for in ways that some people consider significant.

Originally Posted by rkomar

Okay, if the keys are in persistent storage, that makes it easier for now. However, that can change. In my job, I'm working with Trusted Execution Environments that allow stuff to be encrypted and decrypted on the device without the keys being accessible within the main OS, and the environment is becoming more widely available in systems with ARM CPUs. Using the TEE to encrypt the key store would make it hard to crack. You'll probably be back to trying to fish the key out of the app when it has the ebook open (assuming that they don't keep the key in the TEE and do all decryption inside it).

I'm sure somebody will try that, and as soon as they do, it will be cracked by someone opening the book and dumping the unencrypted contents as it gets processed. And as soon as a single person makes an unencrypted copy available, the DRM is moot; there's no need for everyone to break it.

In fact, what you're describing is an awful lot like the way that most hardware copy protection dongles work. Those apps are invariably cracked pretty quickly. To do so, someone literally runs the app and uses every feature. As each page of the encrypted binary gets decrypted by the hardware device, the cracker writes the decrypted code to the same offset in a new file. After every page of code has been decrypted, the person removes the code that does the decryption, and now has a decrypted version of that app. It ends up on warez sites shortly thereafter.