W3C

– DRAFT –
FHIR RDF

02 December 2021

Attendees

Present
David Booth, EricP, Gaurav Vaidya, Gopi, James Champion, Jim Balhoff
Regrets
-
Chair
David Booth
Scribe
dbooth

Meeting minutes

Context generation

james: working w EricP

Issue 76: Ordered lists

https://github.com/w3c/hcls-fhir-rdf/issues/76

jim: Figured out my error, got @list working to round-trip RDF lists using framing.

david: Does the frame need to be different than the @context, or can it be the same?

jim: Mostly the same, but also need to say what node you want at the root.

david: Similar to fhir:nodeRole fhir:treeRoot ?

david: Could it be a part of the standard @context/frame or is it specific to the instance data?

jim: I think it is based on the structure, not the instance data.

gaurav: I think it may need to be separate from the @context.

david: I had the opposite impression, but IDK.

ACTION: Jim to see if framing can be in the @context

ACTION: Jim to try out some round tripping made.

FHIR JSON-LD playground

eric: Made lots of progress: https://ericprud.github.io/fhircat-playground/playground/
… Need to generate lots of @contexts. Made an on-the-fly @context generator, fired by loading the FHIR structure def. Makes the @context production faster.
… Only replicated what we had before, but in process of adding more functionality to FHIR JSON-LD playground.
… Need a github repo for it, cuz it needs to own the github.io.
… Making a library for it.

ACTION: James to check in his FHIR JSON-LD code and compare with eric's code

issue 94, lack of concept URIs for CodableConcepts

gaurav: code system is identified by both system and code. Challenge is to add a prefix to the code, to produce a proper URI. But the prefix is not always the same as the system URI. Looking to see if there is a repo of these codes.

david: system is a namespace for a set of codes

eric: a valueset is a subset of codes from one or more system

eric: NamingSystem is like a namespace document in RDF land.

gaurav: NDC page gives details info about how to make the codes.

gaurav: Both CodeSystem and NamingSystem documents are machine readable.
… They're available from HL7 site.

david: What URI goes into System property in instance data?
… It looks like the NamingSystem is additional documentation about the CodeSystem.
… And it seems to be non-authoritative third-party annotations on the CodeSystem.

gaurav: CodSystem has only one defining URI, but NamingSystems can have multple identifiers.

gaurav: The HL7 terminology website has a registry for FHIR CodeSystems. If we can get the prefixes put in there, that would be a good place for them.

gaurav: To get it in there, maybe needs to be added to a NamingSystem

ACTION: Gaurav to see if prefix could be added to CodeSystem definitions

eric: We postulated that HL7 would publish a URL template for CodeSystem URIs, and the owner would then claim it.

gaurav: If we can convince HL7 to add a prefix field to CodeSystem definition, then we're off to the races.

ADJOURNED

Summary of action items

  1. Jim to see if framing can be in the @context
  2. Jim to try out some round tripping made.
  3. James to check in his FHIR JSON-LD code and compare with eric's code
  4. Gaurav to see if prefix could be added to CodeSystem definitions
Minutes manually created (not a transcript), formatted by scribe.perl version 159 (Fri Nov 5 17:37:14 2021 UTC).

Diagnostics

Succeeded: s/subset/subset of codes from one or more system

No scribenick or scribe found. Guessed: dbooth

Maybe present: david, eric, gaurav, james, jim