Originally Posted by Alexander Turcic
Last month we learned how the anti-piracy outfit BREIN is able to track down
suspected e-book pirates by matching digital watermarks inside e-books with the transaction records they receive from e-book sellers (not without causing a political controversy
). If you've caught yourself wondering how such a digital watermark looks like, a computer science student at the University of Twente was able to get his hands on a watermarked EPUB file and dissected it into pieces
. The result is astonishingly simple: embedded on every page in the book is a minuscule image that turns out to be an ordinary barcode. This barcode contains the transaction code which uniquely identifies the e-book purchase.
So is that really all? The author of the post doesn't exclude the possibility of other "invisible" watermarks, such as random variations in text or punctuation. How to know for sure? By comparing two separate purchases of the same book.
If these books are watermarked, why do they go to the expense of putting DRM on them? Just wondering.