Meeting minutes
R6 priorities
dbooth: Want to prioritize any R6 changes needed
tim: If we have approval for R6 changes, the implementation can be done after publication, but we'll want them in the spec (RDF page, examples, OWL, shex).
… There's a few things I'd like to change, including Concept IRIs on CodeableConcepts.
… Main thing is getting consensus on these changes.
tim: for #223, we create a new class for the element, like syntactic sugar. And we also do it for other complex elements.
… It makes the class hierarchy not match the type hierarchy.
ericp: There are a few that cannot be anonymous, becuase they are referenced elsewhere.
ACTION: EricP to review #223
tim: On #224, it's about how shex and OWL allow use of these underscore classes
tim: Recent issues: w3c/
ericp: This comes up in 3 places: modified elements, modified attributes, and ...
ACTION: Tim to reach out to jim for input on these OWL issues
Concept IRIs for CodeableConcepts
tim: Options for concept iris: w3c/
dbooth: option 3 arc is too broad
dbooth: straw poll
tim: option 1
ericp: Need examples of only Codings
ericJahn: Want to see examples
erich: ditto
ken: abstain
steve: abstain
dbooth: ok, needs more discussion
ericp: Need to see all the places that Coding is used directly.
… Then see what we gain/lose
ACTION: Tim to see all the places that Coding is used directly
Extending the ontology
https://
https://
ken: Taking SULO principles of separating roles and entities
… around concept of Agent and Agent roles
… want guidance on: Unified Care Model (UCM) has foundational alignments with BFO, and structural alignments w SULO. Beginning to document those.
… But want to integrate also with other ontologie: FHIR ont, and linking concepts with that, and the HUD model.
… Want input on that.
… And want to add skos for closematch and exactmatch.
dbooth: want to see an example of how you envision doing it, to help think it through.
tim: You can directly make references using FHIR CodeableConcepts
… That's helpful if it is within FHIR.
ken: Not ready to delve into datatyping and CodeableConcept yet.
… But don't want to restrict that extension.
tim: We can represent these profiles as OWL classes to make more specific mappings: https://
ken: I sometimes get confused about the business problem you're trying to solve.
tim: We're talking about mapping FHIR to another system, within OWL and RDF.
… But FHIR also has a way to do mappings within FHIR.
ericJahn: Should start making some example RDF using CodeableConcept within FHIR and outside of FHIR.
… I think that will help everyone see if it is adaptable to real world scenarios.
ken: Granularity issue. I'm currently focused on the concept, not the binding.
tim: The social determinants of health from Gravity project, has more specific FHIR concepts.
… We can create OWL concepts that correspond to those profiles. I think it would make it easier for you to map to them.
ken: Absolutely. They've done a lot of good work on concepts.
ericJahn: We've never gone beyond: "what happens when there is no patient relation?"
… That's when we need something other than Gravity.
… Want UCM be able to do both.
ken: I can try relating a subject to a condition.
ericJahn: should do two random domains, then CodeableConcept approach
ACTION: Ken and EricJahn to make examples
ADJOURNED