Now, a lot of developers are focusing on smartphones because and they are NOW becoming more and more gaming capable, a few years ago cellphone gaming meant a crappy 2d java game but now... I see this phones as menace to portable gaming consoles, not to consoles in general, that would be exagerating the issue, and even so... I dont want my iphone dead after a couple hours of gaming.
But if you really want a touchscreen device on the cheap there are touch screen kits for most popular notebooks, just search on ebay. There are also touchscreen monitors bellow the 200 bucks, these can be hooked up to the existing desktop in the classroom and serve just right.
I would advice you not to set high hopes for resistive touchscreens, they tend to reduce the lcd brightness and contrast while incresing reflections and diffusing the image, if your really really want this implemented try before you buy.
The google method of digitalizing is at least adequate, it doesn't require much human intervention, is a fast, easy to do process and renders copies of adequate quality, I have been using GB for quite some time and I have yet to found an illegible or missing page.
An open standard for scanned works is nice and everything, but is it within a company interest? how could google, microsoft or whoever is into scanning work benefit from them? wouldn't it be better for THEM to define the standard? why should they bother if they can still make a buck and the costumer/user be happy? In many cases good enough is enough.
Going back to the metadata issue, can you see yourself correcting some metadata? adding the year a book was published into a simple text box? I can clearly see myself doing that, however, can you see yourself scanning a book or two, or a hundred with consistent quality?. The hardest part is getting done, our literate robotic overlords are doing it four us, if google provided a way for users to add that metadata it would be awesome, someone coming close to what google is doing wouldn't hurt either.
On the eighth day, God created FORTRAN.