14:02:15 <RRSAgent> RRSAgent has joined #hcls2
14:02:20 <RRSAgent> logging to https://www.w3.org/2026/05/29-hcls2-irc
14:02:21 <dbooth> rrsagent, make logs public
14:02:26 <dbooth> Meeting: FHIR RDF -- R6
14:02:31 <dbooth> Chair: David Booth
14:03:30 <dbooth> Topic: R6 changes
14:08:30 <dbooth> tim: Getting jira ticket ready
14:08:41 <dbooth> ... Biggest change is Concept IRIs for CodeableConcepts
14:09:11 <dbooth> dbooth: Suggest the examples have comments saying "# Optional"
14:12:26 <dbooth> ... Also suggest one jira ticket for all changes.
14:22:03 <dbooth> ericp: For modularization, could have one top-level OWL file that imports all of the others.
14:28:01 <dbooth> ACTION: Tim to make a W5 issue, and add to Jira ticket to fix it.
14:28:45 <dbooth> AGREED: add a top-level class W5 and top-level object property, for W5 ont.
14:29:30 <dbooth> AGREED: Fix W5 ont as described in issue 234
14:29:52 <dbooth> https://github.com/w3c-cg/hcls-fhir-rdf/issues/234
14:31:52 <dbooth> AGREED: We will not gen the modifier ont, as described in issue 229
14:34:54 <dbooth> ericp: Propose we clarify that using a resource like Observation, the user must be aware of the status.
14:35:16 <dbooth> AGREED: Propose we clarify that using a resource like Observation, the user must be aware of the status.
14:35:30 <dbooth> s/Propose we //
14:37:35 <dbooth> tim: Re issue 194, rendering of bundles.  An example showed a Bundle that has a ref to two different versions of the same Patient.  They were merged in the Turtle, which is wrong.
14:50:26 <dbooth> ACTION: Tim to get clarification on URIs to disambiguate those two different version.
14:52:23 <dbooth> AGREED: Fix the bundle example having two separate Patient versions, according to the clarification that Tim gets on zulip
14:52:59 <dbooth> https://github.com/w3c-cg/hcls-fhir-rdf/issues/234
14:53:07 <dbooth> s/234/194/
14:54:55 <dbooth> Topic: RDF Lists
14:55:16 <dbooth> https://github.com/w3c-cg/hcls-fhir-rdf/issues/233
15:00:29 <dbooth> dbooth: I suggest we allow the R4 style as optional, but keep the RDF List style as required.
15:01:23 <dbooth> ... And allow them both on orderMeaning properties and non-orderMeaning properties.
15:01:51 <dbooth> ... and optionally include the fhir:index
15:02:48 <dbooth> ... But would that approach cause a problem with the property names?
15:04:38 <dbooth> tim: would need to add that to the shex
15:04:54 <dbooth> ... I could make an extended version of the shex to allow that
15:22:55 <dbooth> dbooth: We have a fundamental dichotomy between being OWL friendly or SPARQL friendly out of the box.
15:23:26 <dbooth> ... I would be uncomfortable backtracking on our decision to use RDF List for ordering.
15:25:32 <dbooth> ericp: They previously said "ordering must be retained because it might be important".  But now they are changing to the opposite default, saying "order does not matter EXCEPT if orderMeaning is set"
15:25:45 <dbooth> tim: And it's only set for 6 elements.
15:27:06 <dbooth> ericp: So HAPI "equals" operator code will have to be rewritten.
15:32:36 <dbooth> tim: Shex generator is not always generating the shapes in a deterministic order.
15:33:59 <dbooth> ... I fixed it in my branch, and noticed that you did also.
15:38:29 <dbooth> ericp: Back to design decisions, we could use OLO for those 6 cases.
15:40:00 <dbooth> ... Part of the motivation for using RDF lists was marketing, to look more like the JSON.
15:46:07 <dbooth> dbooth: We need broader input if we are going to change to the OWL-friendly style.
15:46:16 <dbooth> ACTION: dbooth to create jira ticket for R6 changes.
15:46:42 <dbooth> s/changes/changes in issue 176/
15:47:22 <dbooth> Present: EricP, David Booth, Tim Prudhomme
15:47:27 <dbooth> ADJOURNED
15:47:33 <dbooth> rrsagent, draft minutes
15:47:35 <RRSAgent> I have made the request to generate https://www.w3.org/2026/05/29-hcls2-minutes.html dbooth