Meeting minutes
Approval of minutes
Jan 13, 2022 – Issue 94 (concept URIs), FHIR JSON-LD Playground, Turtle serialization
Jan 06, 2022 – Status updates, Valuesets and Codesystems, Turtle pretty printing
Dec 23, 2021 – Introductions, Issue 94, Semantics of FHIR structure definitions, Personal Knowledge Graph
Dec 16, 2021 – Issue 76, FHIR JSON-LD library, FHIR JSON-LD playground, semantics of FHIR structure definitions
Dec 09, 2021 – Introductions; Issues 76, 94; and FHIR JSON-LD playground
Dec 02, 2021 – Issues 76, 94; and FHIR JSON-LD playground
Nov 18, 2021 – Examples of multiple proposed R5 changes
Nov 11, 2021 – Issues 76, 77
Nov 4, 2021 -- Issues 94, 95
Oct 28, 2021 – Issues 77, 93, 76, 69
Oct 21, 2021 – Issues 76, 69
Oct 14, 2021 – Issues 93, 92, 78
Oct 07, 2021 – Issues 94, 69, 92 and 77
Sep 30, 2021 – Issues 77, 93 and 69
Sep 23, 2021 – Review of HL7 WG meeting; Issue 77
Sep 21, 2021 – FHIR RDF Update at HL7 Working Group Meeting
eric: makes a motion to approve the above minutes.
APPROVED: 7/0/0 Minutes above approved.
brian: May WG will be virtual again. dates on the HL7 website, May 9-13.. Meet on Wednesday 11am. Eastern US timezone.
Third panel of FHIR JSON-LD playground
(James gives us a demo)
james: Currently generates nquads, but will add in Turtle
james: IDK what the Framed tab does.
eric: It adds another input text box for the framing.
eric: Our narrative is to add a third window BEFORE the first JSON-LD playground window.
david: the options that eric is add should be applicable between the first and second windows.
eric: The "framed" button should move the middle window to the left and add the framing window next to it.
james: Did you create that turtle serializer?
eric: Yes, basically.
… I used shex validation to order the triples.
RDF lists and OWL
jim: I've collected previous discussions, going back to work on OWL2.
… Feeling pessimistic. OWL experts considered the issue at the time.
… Sometimes might want to use rdf:first as data property, sometimes as object property.
… Maybe consider having a FHIR RDF to OWL transformation, which would only have to change the list vocabulary.
eric: That would allow you create typed lists.
… I think it's ridiculous that we're 10 years into RDF 1.1 and we're now looking at this problem.
… Most interesting use case for FHIR and OWL, is classification use case.
… Harold Solbrig did a nice example.
https://
… If you play that twice in the same data. E.g., person record has an address with a list of contacts.
… E.g., first address must be home address.
… And another list that's CodeableConcept, and see if we can do in OWL what we want to do.
jim: Often in OWL reasoning, i expect some preprocessing before putting anything into a reasoner.
… You kind of need to know all about it first, and then put in the right things.
… Wouldn't be so bad to say that FHIR RDF needs some transformations before applying an OWL reasoner.
ACTION: Jim to start with the Harold's tutorial
Converting data to FHIR RDF
dagmar: Converting synthetic patients to RDF, trying to make them about one patient, to try SPARQL queries. Trying to see how you can use FHIR RDF in other ways.
… About a million synthetic patients, in HL7 FHIR JSON.
… Analyzing how they are linked.
ADJROUNED