Meeting minutes
IRI stems
gaurav: Talking w both Mondo and Radlex about how to represent their URIs in RDF. Proposed a change to Mondo readme file.
… Radlex, they pointed me to their website and said they are checking on my question.
ericp: Mondo was putting curies into strings, right?
gaurav: We didn't discuss that aspect. Should discuss that also.
DICOM
erich: Issue w lack of suport for null values in RDF lists. Current SPARQL WG is only chartered to work on rdf star.
… RDF allows a triple to be left off. Makes sense to me to leave it out of RDF list, but syntax doesn't support it.
ericp: What's the cost of fhir:null if needed?
erich: If it becomes mixed types it explodes out in JSON.
dbooth: I and a portion of the RDF community views it as a design mistake in the original design of RDF to allow gaps in an RDF list, making it not well-formed.
erich: I don't understand that, becuase it works fine in RDF.
… In DICOM JSON they put null. But can't do that in JSON-LD.
dbooth: It's a TBox assertion, not ABox. It's an assertion about the schema, saying that there is a normally a value for that property,
ericp: OWL says lists cannot be changed by inference.
… OWL rejects any attempt to put axioms in an RDF list where the nodes are not bnodes.
erich: The restriction is for OWL compatibility? ericp: Yes. They needed a building block for places to close the world on lists, and OWL uses lists for stating things.
… Needed to make sure you werenn't messing with the foundation for their inference.
erich: Restriction is to make it a closed list? ericp: Yes.
… Is there a feature for open lists?
ericp: There are OLO lists.
dbooth: Not part of the RDF/OWL standards, and no syntax support for it.
erich: Is it legal to manipulate lists in OWL?
ericp: If you have an OWL ont and it has "all distinct" and you ___ OWL will reject it.
ericp: If you leave gaps in lists it will be harder to work w OWL.
… I think the fastest path forward would be a fhir:null
erich: Literals become an issue. How to handle it in the JSON processing? Could say "If you see this, replace it w a null"?
ericp: Could invent jsonld:null value, for use in Turtle.
dbooth; you'd run headlong into the issue that pat hayes.
ericp: We want something to stick in the list, for JSON-LD.
ericp: It needs to show up in triples somehow, such as jsonld:null
ACTION: EricP to follow up on email list
josh: A bnode feels right
ericp: You can't merge two RDF lists.
ADJOURNED