Stream: conformance
Topic: Inheritance and Cardinality Changes
Gino Canessa (Jun 10 2020 at 21:40):
I've been looking through Zulip and haven't found anything: are Resources allowed to change cardinality on inherited fields?
Currently in R5, ConceptMap inherits from CanonicalResource. But CanonicalResource.identifier is 0..* and ConceptMap.identifier is 0..1.
Updating some tooling and it's throwing up because one should be coded as an array and the other should not.
Is this legal? If not, I can file a Jira. (If it is, I may file one anyway, but that's a different issue).
Thanks!
Lloyd McKenzie (Jun 10 2020 at 21:41):
CanonicalResource is a pattern. It's not strict inheritance
Lloyd McKenzie (Jun 10 2020 at 21:41):
From a class perspective, you can say that ConceptMap implements the CanonicalResource interface
Gino Canessa (Jun 10 2020 at 21:41):
So fields (e.g., id, meta, etc.) should not be inherited?
Lloyd McKenzie (Jun 10 2020 at 21:41):
(Though if you pass it more than one identifier, it'll yell)
Lloyd McKenzie (Jun 10 2020 at 21:42):
It has proper inheritance from DomainResource
Gino Canessa (Jun 10 2020 at 21:42):
Where? In R5 it's listed as Type: CanonicalResource
Lloyd McKenzie (Jun 10 2020 at 21:42):
@Grahame Grieve, we probably need to clearly distinguish the difference between 'is a' and 'implements'
Gino Canessa (Jun 10 2020 at 21:44):
Ahh, the base type is listed as deriving from CanonicalResource, but then fields are defined as parented by DomainResource. Ugh.
Gino Canessa (Jun 10 2020 at 21:44):
Thanks though
Gino Canessa (Jun 10 2020 at 21:48):
Hmm.. there's no way for me to know that difference consuming the StructureDef is there? It has "baseDefinition": "http://hl7.org/fhir/StructureDefinition/CanonicalResource",, but just doesn't inherit any fields from it. Do we always just filter that off for CanonicalResource?
Grahame Grieve (Jun 13 2020 at 21:14):
the only way to tell in the StructureDefinitions is that the definition for CanonicalResource has extension http://hl7.org/fhir/StructureDefinition/structuredefinition-interface = true
Gino Canessa (Jun 17 2020 at 19:04):
Created FHIR#27839
Lloyd McKenzie (Mar 27 2022 at 04:21):
I was supposed to write up an algorithm for how code generation to support an interface might work. Please see it here: https://docs.google.com/document/d/1YrdzZoChWCBCClynZ_5jLTXXl0_l0MdL2BdUnSWumh4
Grahame Grieve (Mar 28 2022 at 21:49):
I don't understand how this actually addresses the problem, honestly.
Grahame Grieve (Mar 28 2022 at 21:50):
perhaps you could add some actual code
Last updated: Apr 12 2022 at 19:14 UTC
