- From: Jim Hendler <hendler@cs.umd.edu>
- Date: Wed, 16 Jun 2004 15:32:19 -0400
- To: public-rdf-dawg@w3.org
- Message-Id: <p06110403bcf646e2d316@[10.0.0.11]>
Action: JimH elaborate his proposed use case (RDFS/OWL related) in story form I propose something similar to this, I would ask the editor to help me make it a little more human readable - but wanted to try to be relatively unambiguous (see notes below) 2.x RDFS/OWL schema repository query TiedTogether.com, a new social networking site powered by the popular FOAF ontology, has written a crawler which follows foaf:knows links to determine the publicly available properties of new people it will invite into the network. While processing http://ex.org/~someone/foaf.rdf it finds a rdf:property referring to a URI that it has not previously encountered. The crawler queries a schema server's RDF graph to see if the property's domain(s) and range(s) are ones that it has already encountered, so it can usefully use this new property in future searches. ==== Notes: 1 - This is more or less a real use case, names have been changed to protect the innocent (but look at http://trust.mindswap.org/trustOnt.shtml for an example of the sort of "new property" that we might encounter) 2 - while it is vaguely possible that some sort of inference procedure could be used for determining domains and ranges, I have seen no such use to date - so I hope this will be more palatable to Rob than owl:inverseFunctional or other restriction forms, even though those query graphs would be a bit more interesting. 3 - the existence of such schema servers are predicated in part on the fact that I know of several organizations, including mine, that are building these to host large ontologies -- for example, we will soon have an online, queryable version of the NCI ontology hosted in Kowari available for various kinds of queries -- and we will commit to using it to test the DAWG langauge/protocols as they evolve -- the NCI ontology is huge and it was HTTP-GET'ed from our site about 215 times in May accounting for 5778827 Kbytes of traffic (about 30% of our activity), and we've already had 150 downloads in June (accounting for nearly 40%) -- so we are desperately looking for another way to provide information for users just looking for some specific content from this huge beast. -- Professor James Hendler http://www.cs.umd.edu/users/hendler Director, Semantic Web and Agent Technologies 301-405-2696 Maryland Information and Network Dynamics Lab. 301-405-6707 (Fax) Univ of Maryland, College Park, MD 20742 240-277-3388 (Cell)
Received on Wednesday, 16 June 2004 15:32:22 UTC