Closed sparkji closed 8 years ago
I think your first example is fine. So long as the subject of your triple:
_:cruiseHX203 glbase:hasIdentifierScheme http://purl.org/spar/datacite/doi .
is a URI and not an RDF literal. Then glbase:hasIdentifierScheme
being an Object Property is fine. Am I understanding your question correctly?
A side note: We talked about what to put in the hasIdentifierValue
triple as a subject and I think we agreed to prefix things like DOIs with doi:
so your glbase:hasIdentifierValue
triple would have a subject of "doi:10.7284/901886". See https://github.com/ec-geolink/design/blob/master/data/dataone/canonical-identifiers.md.
It make sense. I just have a little confuse on ObjectProperty and DataProperty. My understanding was ObjectProperty must company with an instance of a specific class matching with the range of ObjectProperty. It is wrong, right?
Thanks for your explanation!
Yes,
ObjectProperties have individuals (class instances) as fillers (i.e., in the object position of the triple)
While DataProperties have datatype values (e.g. "xyz"^^xsd:string) as fillers (in the object position of the triple)
It's somewhat confusing that the "object" in ObjectProperty has a different meaning than "object" as in an RDF triple. The description logic people call them "abstract" and call the DataProperties "concrete", which is a bit less ambiguous but didn't make it into the standard documents.
Pascal.
-----Original Message----- From: Peng [mailto:notifications@github.com] Sent: Tuesday, March 8, 2016 11:33 AM To: ec-geolink/design design@noreply.github.com Subject: Re: [design] How to use Identifier, IdenifierScheme, and IdentifierValue (#89)
It make sense. I just have a little confuse on ObjectProperty and DataProperty. My understanding was ObjectProperty must company with an instance of a specific class matching with the range of ObjectProperty. It is wrong, right?
Thanks for your explanation!
— Reply to this email directly or view it on GitHub https://github.com/ec- geolink/design/issues/89#issuecomment-193851782 . https://github.com/notifications/beacon/AJNbInmcWfRcfN2NR_CjAOXH9J8 HTY7hks5praTKgaJpZM4Hr4VC.gif
Oops! I wrote out my reply above too hastily, "So long as the subject of your triple:" should have read "So long as the object of your triple:". Doesn't look like I confused anything but my apologies.
@sparkji can we close this issue or do you have any other questions? Go ahead and close if you're all set.
Thanks!
IMHO, the names are not confusing. They just go back to a programmer's perspective in which we distinguish between objects (such as a buildings) and data types such as integers for their height. The term 'object' in RDF triples simply (and informally) refers to the position of a resource in a statement. Such an RDF object can be a subject of some other triple or a literal. In other words, semantics is a all about context and thus the term 'object' has different meanings in the realms of programming/modeling and linguistics.
Cheers, Jano
On 03/08/2016 09:14 AM, phitzler wrote:
Yes,
ObjectProperties have individuals (class instances) as fillers (i.e., in the object position of the triple)
While DataProperties have datatype values (e.g. "xyz"^^xsd:string) as fillers (in the object position of the triple)
It's somewhat confusing that the "object" in ObjectProperty has a different meaning than "object" as in an RDF triple. The description logic people call them "abstract" and call the DataProperties "concrete", which is a bit less ambiguous but didn't make it into the standard documents.
Pascal.
-----Original Message----- From: Peng [mailto:notifications@github.com] Sent: Tuesday, March 8, 2016 11:33 AM To: ec-geolink/design design@noreply.github.com Subject: Re: [design] How to use Identifier, IdenifierScheme, and IdentifierValue (#89)
It make sense. I just have a little confuse on ObjectProperty and DataProperty. My understanding was ObjectProperty must company with an instance of a specific class matching with the range of ObjectProperty. It is wrong, right?
Thanks for your explanation!
— Reply to this email directly or view it on GitHub https://github.com/ec- geolink/design/issues/89#issuecomment-193851782 . https://github.com/notifications/beacon/AJNbInmcWfRcfN2NR_CjAOXH9J8 HTY7hks5praTKgaJpZM4Hr4VC.gif
— Reply to this email directly or view it on GitHub https://github.com/ec-geolink/design/issues/89#issuecomment-193874249.
Krzysztof Janowicz
Geography Department, University of California, Santa Barbara 4830 Ellison Hall, Santa Barbara, CA 93106-4060
Email: jano@geog.ucsb.edu Webpage: http://geog.ucsb.edu/~jano/ Semantic Web Journal: http://www.semantic-web-journal.net
These are triples we created for a R2R cruise identifier. http://data.rvdata.us/id/cruise/HX203 glbase:hasIdentifier :cruiseHX203 . :cruiseHX203 rdf:type glbase:Identifier . :cruiseHX203 glbase:hasIdentifierScheme http://purl.org/spar/datacite/doi . :cruiseHX203 glbase:hasIdentifierValue "10.7284/901886" .
But, “IdentifierScheme” is Object Property in glbase, should we create triples like below rather than list above:
http://data.rvdata.us/id/cruise/HX203 glbase:hasIdentifier :cruiseHX203 . :cruiseHX203 rdf:type glbase:Identifier . :cruiseHX203 glbase:hasIdentifierScheme cruiseHX203Identifier . :cruiseHX203 glbase:hasIdentifierValue "10.7284/901886" . cruiseHX203Identifier rdf:type glbase:IdentifierScheme . cruiseHX203Identifier owl:sameAs http://purl.org/spar/datacite/doi .