On 26/02/07, Iris Jastram <ijastram_at_gmail.com> wrote:
> I'm so glad to hear that, Tim! You just made my day.
>
> Iris
>
>
> On 2/26/07, Tim Spalding <tim_at_librarything.com > wrote:
> > Not dead at all. Quite alive. I'll post about it as soon as I can.
> > I've had an offer to host the images—continuously, not as a dump—but
> > can't talk about it until it's all agreed to. The legal issues are
> > surmountable.
> >
> > Tim
Had a thought about this recently. Based on purely unscientific
observation, the available collection of Amazon's cover images tend
towards the recently available and commercially popular books (for
obvious reasons). While daydreaming about the possibility of
converting our legacy holdings to digital format, I wondered whether
the contracts that those libraries that had already signed with
Google, Microsoft, et al to digitize their collections would prevent
them from contributing the cover images for their collections to an
open collection of cover images. If not, well, that would be an
excellent way to get cover images for digitized books. ISBN's wouldn't
be available for anything old enough to be out of copyright, so a Web
service that returned an image based on a fuzzy match of title /
author / pub. date or the like (perhaps with a score indicating the
probability of a match) might be a necessary corollary. I see the wiki
page hints at the possibility of a Web service, and the "recon
project" might refer to the book digitization efforts.
By the way, are we talking strictly about book images, or would this
open database also be open to the possibility of journal covers, or
other media such as album and CD covers, DVD covers, etc? Especially
with the latter, I could see this becoming a very popular service
outside of simply enriching library catalogs (e.g.
http://freecovers.net).
Finally, to avoid this becoming another locked-in set of content (call
me skeptical), I would really love to see the collection (images and
metadata) made available via BitTorrent. This would enable libraries
to spread the load, both by acting as BitTorrent seeds for the
collection and by self-hosting the collection (the ultimate in
caching). "coverthing" (or whatever this ends up getting called)
wouldn't get deluged with http requests, so it could be built with far
less infrastructure. And those interested in adding their own APIs to
the collection would be able to innovate a little more freely.
Dan
Received on Mon Mar 26 2007 - 07:20:24 EDT