Just a small clarification, since some details didn't get out in the primary coverage.
The Metaweb technology that supports Freebase is indeed centralized, but only for computational speed. Distributing queries of real-world complexity over a high-latency network is a very difficult problem.
Danny [Hillis]'s original proposal for the Metaweb in 2001 assumed that it would ultimately be widely distributed. For now, we want to ensure that it is practical and useful until the bigger problems are solved.
Which sounds altogether reasonable, assuming the Web in Metaweb
is the one we all know & love (see
Linked
Data).
Tim O'Reilly got a gratifying number of comments from
semweb types, mostly correcting his misapprehension about the
approach to vocabularies, the most robust coming from
Jim
Hendler.
This exchange made me smile -
Tim O'Reilly :
Denny Vrandecic :Correct me if I'm wrong, but I haven't heard about other initiatives that allow wiki-style editing of the ontology itself, where anyone can modify an existing ontology.
Tim, that's exactly what our wiki is about. You are supposed to change the vocabulary in a wiki-way, just as well as the data itself. Need a new relation? Invent it. Figured out it's wrong? Rename. Want a new category of things? Make it.
Applied mythbusting, great stuff.
Elsewhere,
Shelley
posts on the matter highlighting the difference between the Web and
Wikipedia architectural styles.
Over at Casa O'Reilly, Robert says:
We think the Semantic Wikipedia is great. We'd love to find a way to collaborate.
Robert, if you mail me I'll don my SWEO hat and make the appropriate introductions. (Or you could perhaps just subscribe to linking-open-data).
One thing that's bugging me a little about this is I'm sure I
had some contact with one of the Metaweb folks, maybe a year and
a half ago, but can't find anything in my mail archives. I might
be mixing it up with something else, but I associate the idea of
a database optimised for not using DELETEs with it...dunno where
else I could have got that from.