http://rowling.books/index (served as a JSON or Marc file)
http://rowling.books/hp1.txt (or ePub, mobi or HTML)
It won't happen while publishers keep putting $40 books in the bookshop and expect to be able to sell them. Nor will it happen while Elsevier still charges $80 for that article you can only see the abstract of. But between getting said book as a text file from untold websites or torrents already or for half a dollar used at Amazon or some local second-hand bookshop, the above might really be the answer.
And let Goodreads, IMDB and Amazon worry about the metadata and discovery and the review aspects.
What we have right now flies in the face of Tim Berner-Lee's statement "Cool URIs don't change". They really don't and they never should. And librarians are hating it.
We have a gross inefficiency in the system as it stands, and neither Amazon's Lending Library thing nor paywalls nor keeping on charging for books on a per-book basis are the answer.
We don't just need a Steam for books, we need the web and some way of nixing this whole DRM problem. DRM is a symptom of the problem, and that's problem is getting authors paid. If it's out of my taxes (or some other way that leaves me paying about $10 a month per household or even some reasonably priced subscription service, Netflix-esque) and there's adequate metrics for determining which author should get paid for which publishings of theirs, then I'm all for it.
We've managed to fund public broadcasting. Maybe it's time to fund public booksharing.
I just hope that 2020 will be the year of:
or
http://asimov.books/foundation.html
or
