Stream: fhir/infrastructure-wg
Topic: FHIR Version
Grahame Grieve (Dec 13 2021 at 23:11):
I have started getting R4B ready for publishing. Per the agreement with FMG, I am transiting to using 4.3.0-CIBUILD, etc, so that the rest of the R4B sequence will be 4.3.0
Grahame Grieve (Dec 13 2021 at 23:12):
this raises a question I hadn't considered, and we didn't discuss:
Grahame Grieve (Dec 13 2021 at 23:40):
at this point, the FHIR version enumeration used in StructureDefinitions lists every release of FHIR from the start, including the ballot and interim versions.
Grahame Grieve (Dec 13 2021 at 23:41):
that doesn't have be the case anymore - I could just have 4.3.0 as "R4B sequence". or I could list all the interim versions - 4.3.0-DRAFT
Grahame Grieve (Dec 13 2021 at 23:41):
anyone have any thoughts?
Vassil Peytchev (Dec 14 2021 at 01:02):
Are these enumerations driving how an artifact is referenced in an IG?
Jean Duteau (Dec 14 2021 at 01:12):
Would anyone need to reference a specific interim version? Currently I am building off the balloted r4b but will want to switch to the published version. Thus I can make do with just 4.3.0. But others might want to stick to the balloted version for some time?
Grahame Grieve (Dec 14 2021 at 04:20):
I'm not asking about the package version - that will be 4.3.0-XXX. I'm asking about the enumerations inside the resources themselves
Grahame Grieve (Dec 14 2021 at 04:21):
and no, this is not what drives versioning from for IGs - they will run off the package version. In dependencies. But canonical versioning will run off the resource version not the package version
Grahame Grieve (Dec 14 2021 at 04:21):
for now, I'll be continuing with the practice that the resource versions match the package version - it turns out to be a significant rewrite to change that arrangement
Lloyd McKenzie (Dec 14 2021 at 05:51):
Where are the enumerations used?
Grahame Grieve (Dec 14 2021 at 06:44):
CapabilityStatement.fhirVersion, ImplementationGuide.fhirVersion , StructureDefinition.fhirVersion
Ewout Kramer (Dec 14 2021 at 09:22):
These interim versions sound like release candidates to me, and the intention is that these versions represent 4.3.0 in its final shape, so we could consider them to be 4.3.0, and do not need a persistent label in the version enumeration.
Grahame Grieve (Dec 14 2021 at 09:55):
well, R5b draft published this week will have one, because I'm currently tied to one meaning. But we can consider that. it does mean that you won't be able to tell instances from the candidates apart
Last updated: Apr 12 2022 at 19:14 UTC
