Meeting minutes
Concept IRIs
gaurav: Working on descriptoin document, but stuck on algorithm for percent-encoding the codes.
https://
gaurav: Also need example of a real code that is not just ASCII. Have not found one yet.
… Emailed Martin Durst, IRI spec author, to see if there's an existing algorithm. Also searching the terminology server for non-ascii example.
eric: Not optimistic about finding one in the FHIR term server.
Intros
JohnWalker: tech consulting, linux, FHIR RDF work, round-tripping of FHIR from JSON, JSON-LD to RDF.
Concept IRIs
gaurav: Deepak is going to try to run my code, and look for improvements.
dbooth: Might want to expand the algorithm to handle compound keys, if we end up defining our own.
gaurav: Eric and I discussed the problem of needing an IRI template to put something into the middle of the IRI.
RDF lists
jim: Emailed OWL API list, got a response back.
https://
jim: Described the situation, how RDF lists get dropped from the parsing. He agreed that it's a spec violation, that part of the code is convoluted. Think they would accept a code change if someone contributed it.
… Or we could provide a SPARQL transform to use before loading their data into protege.
Eric: Can you set up a minimal example I can run?
dbooth: Do a straw poll?
eric: AFAIK FHIR profiles don't go beyond the first or second item in a list.
… That's how FHIR profiles express slicing. AFAIK that's the only thing in FIHR that uses ordering.
dbooth: Straw poll: Should R5 go for RDF lists? Brad: IDK; Jim: interested in exploring RDF lists more; Gaurav: also want to explore more, but w time limit; John: want to tranform to RDF; eric: RDF lists would simplify the preprocessing; dbooth: I think it would be worth changing to RDF lists in R5 if we can.
ACTION: Eric and Jim to look at OWL API upgrade possibility
ADJOURNED