Meeting minutes
Topic; Concept IRIs
gaurav: https://
gaurav: Also looking at coding system, to find IRI stems for all of them. Found the publisher field. We cannot always ignore HL7 publisher field, because sometimes they're referencing external vocabs.
… But good for most.
… If I filter out everything that's a draft or retired, that leaves 755 code systems and 424 naming systems.
gaurav: next step may be 1. figure out the number of systems needing IRI stems. That will scope the work.
… Another option: 2. Make larger and large change requests for adding the IRI stems.
… From that approach, maybe we should come up with a short list of the ones most needing IRI stems.
dbooth: Sounds like a good start.
… Maybe ramp up to 10-20 next
fhir:link missing from MedicationStatement.subject #121
eric: The problem is that the root node is a blank node, so it doesn't have a base URL for creating the subject URL.
eric: We could make them relative URLs, so wherever people put them, they'll work.
… Or we could use example.org
eric: If we use relative URL, for a patient, it will become "../Patient/123"
dbooth: example.org seems like a less risky approach.
eric: Examples are not set up to be an imitation FHIR server, because they're all in the same dir.
… But we could put them into their own directories -- 190 directories.
eric: The XML and JSON examples could also be set up that way.
dbooth: Seems like it's worth raising an issue with FHIR in general.
ACTION: DBooth to ask on zoom about putting the examples in subdirectories
datatype for xhtml #119
eric: We decided to adopt the xhtml literal type.
ACTION: Jim and Deepak to implement the xhtml literal type
jim: Will get a PR ready
dbooth: Lets' bundle multiple PRs in to the same Jira ticket
align Reference with canonical
#120
(replaces issue 70)
eric: Inclined toward option 1
dbooth: Should solicit more community input before going with option 1.
… or 2
eric: If I get a bundle, that link is useful.
… One of the observations in the bundle may have an abs url, and another in the bundle may have only a fragment. I want to know if they point to the same thing.
dbooth: Need to first find out if the vertical bar can appear after the hash. That will determine whther it is always an IRI.
ACTION: eric to look for a canonical with a vertical bar and frag ID
ACTION: eric to ask on zulip about hash and vertical bar in a canonical
ADJOURNED