In the process of scanning and storing images of pages from over a million books, we've benefited from the fact that the book is, generally speaking, a relatively uniform medium. Most books are a bound collection of flat pages, and the variations between volumes lie primarily in page size and number of pages. Our scanning processes work well for almost all of these titles -- we scan the images, clean them up, and present them for view on the Web.
But a small subset of books possess peculiarities that make them hard for us to archive. For example, we hit a snag when trying to scan pop-up books, how would one properly present a pop-up book in a browser? We've faced similar problems in scanning and displaying Braille books and flip books.
However, today I'm pleased to let you know that we've made some headway with one type of volume we've struggled with in the past: books employing scratch-and-sniff technology. Using special equipment and tricky JavaScript, we're now able to capture some of the smells during the scanning process and then embed them in your web browser when you preview these titles in Google Book Search.
http://booksearch.blogspot.com/2008/...ls-better.html