Here is a recent review of Sony's new eBook. Forgetting about the price issue for a moment, the biggest drawbacks of eBooks in general has always been the flexibility and readability of the hardware. If you can't read them comfortably over long periods of time (many people read books for an hour or more straight), they will not replace paper books.
So how should you test an eBook? I suspect that the reviewer at Tech Review (linked above) read it sitting in a chair in his office or home. Where do you read? Personally, I lie down on my back on my couch or on a lounge chair by the pool. And I prefer magazines and soft paperbacks so I can fold them up into smaller footprints (so I don't block the sun by the pool). If I tried to do this with an eBook, I wonder if it would be as easy to read as the reviewer found in his test?????
This is a great example of what call Activity-Based Evaluation. You always need to design realistic tasks to test a device like this. I don't know what variety of environments and postures people like to read in, but a good test will have the most common, most important, and most likely to cause problems. If Sony's eBook can be successful in all of these (and the price comes down), I would be a likely customer for them.