Textual watermarking from Amazon : United States Patent: 7610382

Oh, I can imagine how the author is going to love this. Look at the phrase "... may not significantly alter the meaning of the excerpt ...".

Tell the author that all the time spent on finding the exact turn of phrase was wasted; don't bother with a thesaurus to find which word you want, because Amazon can just replace it to anything that approximately matches.

To be or not to have presence, that is the controversy.

Yuck!

As someone who likes actually buying and owning things, I'm getting sick of legitimately purchased content being worse than the pirated alternative.

This would be disastrous for scholarship. While I don't know anybody using the Kindle right now to conduct research, it's not too hard to imagine that this could become commonplace in the future. I already use Amazon's Look Inside feature when I need to refer to a section of a book. But if they were to implement this patent, then I can't rely on any quotes from a source as being accurate.

Imagine now that you have cited some text that Amazon silently munged and included it in your own book, which is now, of course, offered for sale via Amazon. The original text could now be doubly harmed!

I first saw this on slashdot:

http://yro.slashdot.org/story/09/10/28/2236235/Amazon-Patent...

To me it looks like there would be ways to circumvent it. It also looks more effective the larger the body of content, so aimed at large scraping projects. Might not be conclusive on a single article.

I might have prior art on this, circa 2002; the original code didn't survive but I have an online posting detailing it.

An ebook water-marking script in Perl which edited the book meta-data (index, contents, preface, indentation, position of page numbers, fonts sizes, colophon and other stuff to generate unique "on demand" copies of texts that could identify the buyer.

The claims read a lot like a system described in a Tom Clancy novel published in the late 80s.

That system was used to discover the source of leaked classified materials. It generated unique versions for each recipient by substituting equivalent text.

I'd suspect they are looking into this for their reviews. There are many, many websites that screenscrape the amazon reviews, and try to make you think they were from that site instead.