So the draft of the Epub 3 spec came out a few days ago, and it has a number oof really shiny features. It's based on HTML5, and there's one part of HTML5 in particular that has me drooling. It's a tag called canvas. You can use this tag to draw content when you read the ebook.
So what can you do with it? I can think of 2, but there are probably a bunchaton.
First, think about what this means for graphs. With the current Epub spec, if you wanted a graph in a book you embedded an image. But with Epub3, you can create the graph on the fly from a set of numbers. What's even better is that eventually we'll get the ability in reading apps to edit the content; this means you'll be able to input the data for the graph and your reading app will draw it for you. Think of what that could mean for textbooks.
My other idea involves graphic novels. Right now there are several propreitary comics formats as well as CBR and CBZ. The latter 2 are simply zip and rar files filled with images, but the propreitary formats do a lot more. they have code that can control how you see the page images: pan, zoom, rotate, etc. I'm pretty sure that Epub3 should be able to support code similar to the propreitary formats, and it will use the canvas tag to accomplish it.
You might not know much about the canvas tag so I found a code library that shows off a few tricks. It's called, and even though it's still in early development I think it's worth looking at. It will show you a few thimgs you can do with this tag. of course, I don't think you'll be able to use this code library in your ebooks, but ti' still fun to look at.
I'd embed an example here, but my WordPress install doesn't like HTML5 tags.