[LazyScot]

Originally Posted by zelda_pinwheel

this is a valid point (the people who say "if i *can* do it, it must be legal", simply because they don't know any better) but i still think it would be much better to try to act against this with an effort of education rather than ineffective, pointless, costly and very annoying (to legitimate customers only !) drm.

I agree that education and no DRM is the better way, but I'm not convinced it will succeed 100%. The question is whether or not it will succeed enough to allow a DRM-free model for paid-for content to work. I'd much rather not have to worry about having keys for my house and car, but I don't think the "education" system protects them; in the same way the "honesty box" is not a widespread system (though I agree it works in some cases).

Why is DRM *necessarily* costly?

Originally Posted by zelda_pinwheel

yes, "draconian" seems apt. but can "a more reasonable drm scheme" exist ? what would it look like ? how would it work ? perhaps a social drm / watermarking scheme, but even then, as i said, i have some reserves simply because you never know what could happen which is beyond your control. i would be interested to hear some ideas or concrete suggestions though. perhaps there is something obvious which i haven't thought of.

I don't see why it couldn't be made to exist, especially if every device had sensible communications, and a built in "move" button for content. It should also be possible to make it completely transparent and cope with device loss/failure (assuming some sort of content escrow agency could be set up).

Originally Posted by HarryT

Even in that circumstance, all you could do would be to try and encrypt the original file with every possible key until you got a match for the encrypted file. For a 16-digit credit card number, such a "brute force" approach would probably be impractical.

I'd guess that you could reduce the CC numberspace quite substantially, say to 8 digits, based on knowledge of likely geography, card suppliest, likely number creation dates, the actual number formats, etc. 10^8 should be brute force searchable over a weekend for a reasonably powerful computer, assuming reasonably small block sizes within the file. And a simple syntax check of what comes out should allow you to decide if the key is probably right, even without the original file to hand. (Though I suspect it is easier to just go to suitable sites and buy the numbers you want, if a recent New Scientist article is anything to go by.)

Originally Posted by zelda_pinwheel

it depends what you mean by "distribute" ! i wouldn't put it on a filesharing site, but that's a decision independent of any watermark ! i *would* share it with friends / family, however, it's true, i'd feel less comfortable than with a non-marked file, because once again, it could be lost / stolen ; or my trustworthy friend could share it with just one of her trustworthy friends who could share it with a friend who turns out not to be trustworthy at all (and who doesn't care, because he doesn't know me, after all...), and then my name is on a file on a p2p site and i am not happy about that. but i still want to be able to share with my (immediate) friends, and not have to worry about such unlikely but possible hypotheses. this is (one of the reasons) why i'm against drm and why i'm not even too convinced i accept social drm, because of possible circumstances out of your control which will have direct repercussions on *you* even if you are not directly at fault.

Interestingly, I think your case could be used to argue for DRM (sorry). It would either prevent the case being able to happen, or provide a fingerprint trace of who was actually responsible.