Freebase looks “lite” enough to actually be useful.
The problem of matching up the Freebase “unique” identifiers with the identifiers actually used by people in real communications remains.
Another problem is how to enable people work in a highly distributed fashion in terms of authoring matches between identifiers.
Are you using Freebase?
Comments?
Hello Patrick
I have been investigating Freebase API (http://www.freebase.com/docs/topic_api) several times in order to write a Topic Maps extractor for it but never actually started the work. Freebase has it’s own identifier space but you can map Freebase identifiers to OpenCYC identifiers, for example, using a service such as AlchemyAPI (http://www.alchemyapi.com/api/entity/urls.html). Wandora already supports AlchemyAPI (http://www.wandora.org/wandora/wiki/index.php?title=AlchemyAPI_extractors).
Freebase is also available as a single data dump at Infochimps (http://infochimps.org/datasets/freebase-data-dump).
Kind Regards,
Aki
Comment by Aki — August 4, 2010 @ 1:07 pm
Aki,
Thanks!
The data dump sounds interesting but a virtual topic map approach a la Robert Barta would be more interesting, at least to me.
Mostly because I don’t think there is enough storage space to copy all existing data over into topic map format. Not to mention that conversion is always iffy at best.
A virtual topic map approach enables data to safely maintain its current residence and yet be reliably combined with other data, which also maintains its current location.
Comment by Patrick Durusau — August 4, 2010 @ 6:43 pm