I’ve written up the process used by infovore to create :BaseKB from the
Freebase Quad Dump here
https://raw.github.com/paulhoule/infovore/master/docs/decoding_the_freebase_quad_dump.txt
I’m doing this now not just to demonstrate the correctness of :BaseKB, but also
to demonstrate the correctness of the Freebase Quad Dump on which it is based.
I’ve got concerns that other possible export formats might be “valid” RDF but
may not maintain the properties that make SPARQL queries against :BaseKB
function almost exactly like MQL queries against graphd.
Infovore, the framework that creates :BaseKB, is available on github
https://github.com/paulhoule/infovore
this, plus the above documentation, make it possible for anyone to verify these
claims. Infovore contains a test suite that can be runs SPARQL queries against
a triple store loaded with :BaseKB that confirms correct operation. Infovore
passes all tests when run against the 2012-11-04 quad dump.
A 1.0 release of infovore is in progress. This is a matter of a single patch to
locate some temporary files in the right place and a small amount of additional
documentation. It may take a few more days because a complete test cycle
including loading into a triple store and running tests takes about 12 hours of
wallclock time.
People are quite aware of the value of testing of software, but it’s taken
longer for people to realize that data products need compatibility testing,
particularly in the RDF and semantic space.
I’d like to advise Freebase to resume publication of the quad dump until it can
demonstrate the correctness of any alternative data export. In fact, with
infovore available under a Apache License and all of my claims independently
verifiable, the freebase quad dump could remain in use indefinitely.
Freebase users should demand a correct export.
------------------------------------------------------------------------------
Monitor your physical, virtual and cloud infrastructure from a single
web console. Get in-depth insight into apps, servers, databases, vmware,
SAP, cloud infrastructure, etc. Download 30-day Free Trial.
Pricing starts from $795 for 25 servers or applications!
http://p.sf.net/sfu/zoho_dev2dev_nov
_______________________________________________
Dbpedia-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
