Stream: IG creation
Topic: IG fhirVersion
Elliot Silver (Jan 08 2022 at 17:28):
What is best practice for filling out an IG's fhirVersion? Does this differ for profiles? It looks like this element needs to be filled out to the patch level, so, how can I say an IG or profile works with R4 rather than 4.0.1? I can obviously fill in 4.0.0, and 4.0.1, but that doesn't help me when 4.0.2 or R4C is released (and with which it should be compatible, assuming FHIR follows it's compatibility rules).
I'm confused by the element definition as well:
this is the formal version of the specification, without the revision number, e.g. [publication].[major].[minor]
I thought [minor] was the revision number. What does a version of the spec look like with a revision number?
Also,
A StructureDefinition does not need to specify the target it applies to as StructureDefinitions will often be valid across multiple versions of FHIR. FHIR tooling can determine whether a StructureDefinition is consistent with a particular StructureDefinition if desired.
Does this imply that for profiles, it may be best to omit the fhirVersion in the SD? (It can't be omitted in the ImplementationGuide due to cardinality constraints, and an invariant.)
Grahame Grieve (Jan 08 2022 at 20:08):
I thought [minor] was the revision number. What does a version of the spec look like with a revision number?
oh that language needs alignment. Can you make a task?
4.0 instead of 4.0.1 - that's the intent
Grahame Grieve (Jan 08 2022 at 20:09):
Does this imply that for profiles, it may be best to omit the fhirVersion in the SD?
not sure whether the tools will ne ok with that
Elliot Silver (Jan 08 2022 at 20:22):
Actually it's worse than "leave off the revision." Is 4 the publication or the major version? I've never thought of the .1 in 4.1 as a major release. And since the FHIRVersion ValueSet is required, can I even do your suggestion of "4.0"?
Grahame Grieve (Jan 08 2022 at 20:40):
actually, no, not in that case
Last updated: Apr 12 2022 at 19:14 UTC
