On Wed, Sep 22, 2010 at 8:26 PM, Ariel Backenroth <[email protected]> wrote: >On Wed, Sep 22, 2010 at 7:46 PM, Tom Morris <[email protected]> wrote:.> On >Wed, Sep 22, 2010 at 7:21 PM, Alex Stinson <[email protected]> wrote: >> >>> I am an active member on Wikipedia and I ran across Open Library when your >> new bot began adding links to our pages( to check out your bot >>> see http://en.wikipedia.org/wiki/Wikipedia:Bots/Requests_for_approval/OpenlibraryBot). >> >> That's interesting. I wonder how the bot figures out which >> OpenLibrary authors go with with Wikipedia pages. >> >> The github project has nothing that shows how that correspondence is >> derived (just a tiny example data file that was spit out from whatever >> process did the calculations). >> >> It'll be interesting to see how closely the OpenLibrary results match >> up with what Freebase did earlier. >> > Tom: The initial corresponding ids for the trial came from freebase and they > match almost precisely to the data in freebase.
Ahh, ok. How are you handling the attribution requirement of Freebase's license? I can't imagine the Wikipedians being terrible keen on lots of links back to Freebase, but it's good that you're paving the way to figuring out how to make the data flow bidirectionally. > However, freebase loaded open library before open library had a work type so > many of their book ids had to be redirected to the work instead of an > arbitrary edition. It would be great to fix up that data on freebase and > i'll get in touch with them about that. Yeah, I'm not really sure why they did that. I guess they figured that some link at the work level was better than none at all. It'd be good to get those converted to real works links, but I'll be curious to see how closely the idea of a Freebase "book" corresponds with an Open Library "work." I think you'll again run into a situation where it's a one-to-multiple mapping because of differences in the way translations are handled, etc. > The query used in case you're interested is here: http://tinyurl.com/2wkxtqh. That doesn't look like it'll handle all the duplicate author and book records which have been merged on the Freebase side of the house since it's only grabbing the first OpenLibrary key (or has all the Freebase author dupe information been integrated into OpenLibrary now?). Thanks for providing more details on the project. It'll be interesting to see how it develops. Tom _______________________________________________ Ol-discuss mailing list [email protected] http://mail.archive.org/cgi-bin/mailman/listinfo/ol-discuss To unsubscribe from this mailing list, send email to [email protected]
