DanC_lap: a user interface to the URI scheme dispatch table in OS X
DanC_lap: e.g. irc: URIs
DanC_lap: worked 1st time for me!
DanC_lap: e.g. irc: URIs
DanC_lap: worked 1st time for me!
jeen: new features include a Graph API, a native disk storage backend, improved unicode support, improved bnode handling,....
jeen: Release Notes and Changelog
jeen: Release Notes and Changelog
dajobe: topics include "Open Data"
dajobe: and focusing on browser tech, web tech
dajobe: (formerly XML Europe)
dajobe: semantic web, rdf stuff seems very relevant
dajobe: and focusing on browser tech, web tech
dajobe: (formerly XML Europe)
dajobe: semantic web, rdf stuff seems very relevant
dajobe: and what about WSDL1?
danbri: Thomas Roessler on DAWG/SPARQL use cases
danbri: There's always a use case for making languages more expressive, and some of these scenarios are close to FOAF concerns too.
danbri: But I wonder whether some of this could be worked around by posing simpler queries to a service known to respect OWL semantics, and couching the OR and negation stuff in terms of classes, disjointness etc.
danbri: SPARQL generally steers clear of that (ie. OWL), but the machinery is certainly similar in some ways. I guess OWL doesn't have facility to represent occurance of data items within some bunch of RDF, it's more in terms of observations about the world.
danbri: Related thought: anyone written an OWL class description -to- SPARQL query compiler yet?
danbri: By this I mean, take OWL class descriptions such as "A ubris:BristolUniversityPerson is a foaf:Person whose foaf:workplaceHomepage is the thing with URI http://www.bris.ac.uk/" and generate: SELECT ?x where ?x type foaf:Person, ?x foaf:workplaceHomepage http://www.bris.ac.uk/
danbri: ...or whatever the proper SPARQL notation is. Actually you could also generate: CONSTRUCT ?x rdf:type ubris:BristolUniversityPerson too.
danbri: I think such exercises are useful, even if not for practical deployment, as they teach us about the related expressiveness of W3C specs. Which are all connected in potential scope more than their formal dependencies show.
danbri: nearby: I've been revisiting my notes on SPARQL as sent to DAWG comments list.
danbri: There's always a use case for making languages more expressive, and some of these scenarios are close to FOAF concerns too.
danbri: But I wonder whether some of this could be worked around by posing simpler queries to a service known to respect OWL semantics, and couching the OR and negation stuff in terms of classes, disjointness etc.
danbri: SPARQL generally steers clear of that (ie. OWL), but the machinery is certainly similar in some ways. I guess OWL doesn't have facility to represent occurance of data items within some bunch of RDF, it's more in terms of observations about the world.
danbri: Related thought: anyone written an OWL class description -to- SPARQL query compiler yet?
danbri: By this I mean, take OWL class descriptions such as "A ubris:BristolUniversityPerson is a foaf:Person whose foaf:workplaceHomepage is the thing with URI http://www.bris.ac.uk/" and generate: SELECT ?x where ?x type foaf:Person, ?x foaf:workplaceHomepage http://www.bris.ac.uk/
danbri: ...or whatever the proper SPARQL notation is. Actually you could also generate: CONSTRUCT ?x rdf:type ubris:BristolUniversityPerson too.
danbri: I think such exercises are useful, even if not for practical deployment, as they teach us about the related expressiveness of W3C specs. Which are all connected in potential scope more than their formal dependencies show.
danbri: nearby: I've been revisiting my notes on SPARQL as sent to DAWG comments list.
DanC_lap: wanted to update this to current SPARQL editor's draft
DanC_lap: but failed to get larch working on this powerbook
DanC_lap: looked around at other tools
DanC_lap: got coq installed, but failed to grok the tutorial
DanC_lap: got alloy (3.0 beta) installed and working, but didn't make much progress
DanC_lap: cf earlier alloy notes