The line between book and Internet will disappear
The secret among those who have poked around EPUB, the open specification for ebooks, is that an .epub file is really just a website, written in XHTML, with a few special characteristics, and wrapped up. It's wrapped up so that it is self-contained (like a book! between covers!), so that it doesn't appear to be a website, and so that it's harder to do the things with an ebook that one expects to be able to do with a website. EPUB is really a way to build a website without letting readers or publishers know it. But everything exists within the EPUB spec already to make the next obvious -- but frightening -- step: let books live properly within the Internet, along with websites, databases, blogs, Twitter, map systems, and applications.
Hugh McGuire on the future of books and publishers as API providers.