People who pay attention to publishing and e-books may have heard the news that German researchers at the Fraunhofer Institute (the birthplace of the MP3) recently unveiled a text-watermarking scheme called "SiDiM" that creates individually indentifiable versions of e-books by making small, randomized text substitutions throughout the book. Basically, the scheme slightly changes punctuation and/or substitutes synonyms, so that each sold copy of an e-book is effectively unique.
Development of the scheme was supported by the German Booksellers Association, and funded by the Federal Ministry for Education and Research, and its hope is that the creation of these “individualized” documents will deter illegal file-sharing, as copies, should they be found on pirate sites, could be traced.
This kind of watermarking would have limited application, of course. To be effective, the variances would have to be chosen at random from the entire text, which means that you could never really tell which text would be altered. Needless to say, it couldn’t be effectively applied to any kind of mathematics, science, reference, or nonfiction work—you don’t want a complex formula, a dose of chemotherapy medicine, or the orbit of Jupiter rephrased by a blind and foolish algorithm.
Shortcomings aside, this kind of DRM is hardly a breakthrough. I first encountered this proposal in the late 1980s, when it came up in a message board on the Science Fiction Round Table on GEnie, an online messaging service. I remember at first thinking that it sounded very clever, until someone pointed out that all it would take to de-identify a text would be to find two or more copies, compare them, make note of the differences, and randomly vary them. That’s easily done—a simple text comparison has been a largely solved problem since the advent of the Unix “diff” command, developed in the early 1970s.
But the fact that the basis behind this security measure was countered 25 years ago by employing a simple tool that’s getting into its 40s is not the silliest part of this supposed new DRM breakthrough. No, the silliest part is the idea that knowing who an e-book was sold to can actually serve as an effective means of fighting piracy. That belief rests on the idea that if you know to whom a file was sold, you can somehow take it out of their hide if you find lots of copies floating around on the Internet.
That, too, is hardly a new concept. For some time now, aware that consumers hate the use of restrictive DRM on their e-books, publishers have been experimenting with something they call “social DRM.” Social DRM is another name for an unencrypted e-book that has the purchaser’s name or some other kind of identifier inserted into it. The theory is that users will be reluctant to share their e-books if they fear legal reprisals or the shame of being identified as patient zero in a widespread outbreak of unauthorized copies.
Last year, Pottermore, J.K. Rowling’s site for all things Harry Potter, employed watermarking on its Harry Potter e-books, what it called “light touch” DRM.
The idea that copyright owners might convince a judge, or, worse, a jury that because they found a copy of an e-book on the Pirate Bay originally sold to me they can then hold me responsible or civilly liable is almost certainly wrong, as a matter of law. At the very least, it’s a long shot and a stupid legal bet. After all, it’s not illegal to lose your computer. It’s not illegal to have it stolen or hacked. It’s not illegal to throw away your computer or your hard drive. In many places, it’s not illegal to give away your e-books, or to loan them. In some places, it’s not illegal to sell your e-books.
So at best, this new “breakthrough” DRM scheme will be ineffective. But worse, what makes anyone think this kind of implicit fear of reprisal embedded within one’s digital library is acceptable, or, for that matter, preferable to old-school DRM?
At this point, the e-book industry, like all parts of the entertainment industry, would be better served by focusing on carrots for customers who opt into buying, not sticks for potential customers who opt out. Rewarding customers, not threatening them, is the most effective tool for getting money in the bank, despite nonsensical revenge fantasies about busting pirates.