Update, March 4, 2010: Penguin Group is well on its way to implementing the future of eBooks that I described in this post. John Makinson, Chairman and CEO of Pearson's Penguin Group spoke at the Financial Times' Digital Media & Broadcasting Conference in London on March 2nd. He showed some concepts for how Penguin plans to transform books into interactive applications for Apple's iPad.
Not a lot has changed in the structure of printed books since the time of Gutenberg. A typical book has a title page, a table of contents, a forward and/or introduction, a set of chapters, and sometimes, an index. Books were organized the way they were because it made it relatively easy for readers to navigate through the content. Today, most eBooks follow the same design and organization as print books, many even going so far as to precisely duplicate the layout of the original print book. However, an eBook is not a print book, and there's absolutely no reason why it needs to be interacted with in the same way as a print book.
In the simplest possible terms, a book is a container for information. The conventions that have developed over the centuries for organizing that information were appropriate for a print medium, but once a book is digitized, the ability to access and utilize the information contained in the book increases exponentially. For example, why bother with an index when a search engine will let you find anything in the book faster and more easily? If you can search for topics and relationships, why do you need a table of contents? Why bother to organize the book into pages when there's no standard page size for eBook readers and text will reflow according to the display or window size being used?
Is there a reason for a book designer to sweat over the details of specific typefaces and sizes when the user can change them with a push of a button or a mouse click? Instead of the page references used by writers and teachers to direct readers to a particular place in a book, why not use named anchors and hyperlinks to identify and navigate to important "landing zones"?
Once a print book is digitized, it can be incorporated into databases and searched along with other documents and information sources. Why limit the reader to just the information contained in the eBook in question? If the reader enters a query and the search engine finds better (or more recent, or more detailed) information from a source other than the eBook itself, shouldn't it point the reader to that information?
If an eBook can be read on a computer that supports audio and video playback and interactive content, why should the eBook itself be limited to text and static images? Replace the static illustration of a physics experiment with a dynamic animation that readers can interact with. Instead of a color image of an elephant, show a video of elephants in the wild. Instead of simply including Walt Whitman's poem "America", link to an audio recording from 1890 of Whitman himself reading the first four lines of the poem.
In short, a book is a container, not the information itself. eBooks limit their potential if they're nothing more than an electronic facsimile of the original print book. The contents of a print book should be the starting point for an eBook, not the end point.