Combating piracy is a balance of making it expensive/difficult/inconvenient for the hacker versus making the document usable. At one level, if the book can be read on a screen then it can be photographed and then re-assembled as the book. The result will not look pretty, and will lack much of the original value (ability to use bookmarks, ability to search and so on), but it is not possible to stop that.
A major potential weakness in most protection strategies is allowing the user to make printed copies (1 is enough). Providing they don’t mind the expense they could then print it out and then scan it back – unless you have added personalised watermarks (you can do this with the image on the screen as well) that identify the source of the copy. LockLizard provide both text (dynamic) and graphic watermark capabilities to help our publishers resist the inclination of end users to make illegal copies. Again, we are looking at a balance of forces. You will want to deliver a high quality result (similar to if not identical with the printed copy?), and adding a watermark will detract visually from your product. At the same time, remember copy shops offered a service where they would take a book and make a scanned copy of it (even making it into a pdf), and any techie can do this with a sharp knife and a lot of patience, to any book in the library. But you can make copying your products less attractive by including watermarks. Obviously you should include the user’s name/email so that they are personally linked to the document (dynamic text watermark) and you should make sure that the watermark is the same color as the main text, and that it is not a header or footer. The reason for this is if it is a rare color a graphics package such as Paintshop Pro or, depending on your pockets, Photoshop, can be set up to automatically remove that color from every image, and also crop off watermarks outside the main text body. So a watermark in the middle of the page using the same color as the text is very difficult to remove. Also, choosing a very thin font, and using a large font size, so that it removing it will damage of lot of text, makes it very difficult to remove the watermark manually, and your average pirate will soon become discouraged. Obviously the larger the document the more work we want to force the attacker to do – ultimately if it is a short document it could be typed in again by hand, and then you are down to how much it would cost to hire bodies to do the grunt work.
Note that scanning printed documents back in will lack much of the original value (ability to use bookmarks, ability to search and so on) since the scanned document is now a bitmap image rather than a working PDF document.