Stream: implementers
Topic: Is this a choice type?
Nick George (May 25 2019 at 00:30):
http://hl7.org/fhir/extension-relative-date.html has kind of event field. It looks like it has two options of type but is not coded as a choice type
Nick George (May 25 2019 at 00:31):
what's the story there?
Nick George (May 25 2019 at 00:36):
ditto for http://hl7.org/fhir/extension-translation.html
Grahame Grieve (May 25 2019 at 01:25):
well, it is, but it's the type of the extension, which is a step away
Nick George (May 25 2019 at 02:07):
yeah ok that makes sense. How about this: do you consider it problematic that http://hl7.org/fhir/extension-workflow-relatedartifact.json.html has the same "name" value as the datatype?
Grahame Grieve (May 25 2019 at 02:41):
i'm not seeing the bit where the name is the same
Nick George (May 25 2019 at 02:50):
well, for the extension
"name" : "relatedArtifact",
Nick George (May 25 2019 at 02:51):
I guess technically the datatype has "RelatedArtifact", which is the same except for case
Grahame Grieve (May 25 2019 at 02:58):
I'm never sure what to do with that name - it doesn't really have any meaning
Nick George (May 25 2019 at 03:03):
Spec calls it the "computer friendly name" for the type - we use it for type name in codegen, so it's problematic when there are collisions - requires a hardcoded renaming
Nick George (May 25 2019 at 03:04):
which isn't that big of a deal, I'm just curious if the expectation is that the name should be unique within a package/logical unit, in which case that would be a bug
Grahame Grieve (May 25 2019 at 03:04):
we haven't said that the name has to be unique. the id definitely has to, but not the name
Nick George (May 25 2019 at 03:05):
fair enough
Lloyd McKenzie (May 25 2019 at 05:25):
It's probably reasonable to at least treat duplicate names in the same IG as a warning
Last updated: Apr 12 2022 at 19:14 UTC
