I got to see the new Sony ebook reader at Borders. It's very nice - really readable, and a lot thinner than I expected. The pages actually look like real book pages. No aliasing, easy on the eyes. And even the background color of the screen looks like the medium-gray of a paperback book.
However, I don't think it will take off in the general public. For one, when you flip a page the wait is "long", 2-3 seconds, which is long for the average person when compared to flipping the page of a book, and more importantly the screen flashes black/white 2 or 3 times each page flip. I think early adopters like us may overlook that, most users will not.
Not to mention the price gouging on accessories. The Device itself is $350, then you have to pay $50 for the dock and $30 for a case - bringing it up to a possible $430 off the bat for the reader and cover.
The screen isn't touch sensitive, which is a little annoying, but not a deal breaker. I'd really like to be able to highlight and mark actual text. I used to do this on an iPaq with the Microsoft reader, which was also very crisp, and legible for reading books. Afterwards I could export a text file of all the passages I had highlighted. But size, cost, usability perhaps wasn't there for this first version of the eBook reader to have touch screen. You can mark a page, which makes a little 'dog ear' icon in the top-right corner of the page and then go back to view your marked pages from the book's table of contents.
The buttons are a little weird, there are 2 sets of "page" buttons. And the menu button is odd, diagonal push below a directional joystick. The joystick doesn't seem to serve that much purpose as most navigation is just up/down. There is also a row of number buttons below the screen, from 1 to 0 (yeah it's "computery" as they did 1-9, then 0. Why not just make it 10?).
Also, the view should be rotatable, but again, not a huge deal.
You can upload images, which are viewable in a grayscale, kind of pixelated/newspapery look. None of the demo books of text had embedded images, so I wonder how that works out. I've seen demo videos that show reading a page of text and having an embedded video (hrm, YouTube on your portable ebook reader)
What I'd *really* like to see, from a geek perspective, is an SDK for developing my own text/software so I can, say, make a dynamic map ;)
I like it, I may consider getting one, but will probably wait for v2.