Saw a presentation yesterday from Dan Reetz who's behind the DIY Book Scanner Project.

He made a real interesting comment which stuck with me.  Computers are going to be the ones reading the most books in the future.  (or something like that, apologies).

If computers will be the biggest readers (because of search), what would books written expressly for computers look like?  I'm not talking metadata, I'm thinking body text.  How should content be expressed to facilitate computer "understanding", or "satisfy" a computer readers.  Would it just be code?