FHIR Chat · R4B versioning · implementers

Stream: implementers

Topic: R4B versioning


view this post on Zulip Lee Surprenant (Jan 10 2022 at 23:20):

I'm looking at https://build.fhir.org/branches/R4B/versioning.html#mp-version (and https://build.fhir.org/versioning.html#mp-version ) and wondering if there's an existing task for updating that (e.g. adding 4.3 for R4B).

Also, now that we have canonical references, is having the profile with the version in the URL still the recommended way for a "naked" resource instance to claim conformance to a particular fhir version?

http://hl7.org/fhir/4.0/StructureDefinition/Patient instead of http://hl7.org/fhir/StructureDefinition/Patient|4.0?

http://hl7.org/fhir/4.3/StructureDefinition/Patient instead of http://hl7.org/fhir/StructureDefinition/Patient|4.3?

view this post on Zulip Lloyd McKenzie (Jan 11 2022 at 04:31):

@Grahame Grieve

view this post on Zulip Lee Surprenant (Jan 11 2022 at 18:08):

given the canonical url for the StructureDefinition of Patient is http://hl7.org/fhir/StructureDefinition/Patient and its version for R4b will be 4.3.0, I think I'd argue that the proper canonical reference would actually be http://hl7.org/fhir/StructureDefinition/Patient|4.3.0

I think there was a discussion somewhere about whether servers should be able to resolve 4.3 to 4.3.0 on canonical references, but I'm not sure where that one landed...

view this post on Zulip Lloyd McKenzie (Jan 11 2022 at 21:17):

There's specific guidance in the spec here: http://build.fhir.org/versioning.html#mp-version. It's informative, so theoretically we could change it, but it would certainly confuse people/break certain implementations.

view this post on Zulip Lee Surprenant (Jan 12 2022 at 01:22):

Yeah, thats the exact page I was looking at. I don't really understand how this whole page could be "informative" given that I think this is the page that defines the fhirVersion mime type parameter; isn't that either normative or on a track to become normative?

view this post on Zulip Lee Surprenant (Jan 12 2022 at 01:27):

To me, it just looks like this page has gone stale and could use some updating. At least:

  1. removal of "(this version)" and "(once published)" language from the R4 row in the table at http://build.fhir.org/versioning.html#mp-version
  2. addition of an R4B row with a version of 4.3

And, preferably (for me at least)

  1. update 2.8.3 to make it clear that the canonical references to core spec StructureDefinition resources work the same as canonical references to any other kind of definitional resource.

I guess I'll open that as 2 separate issues.

view this post on Zulip Lee Surprenant (Jan 12 2022 at 02:14):

opened FHIR#35779 and FHIR#35780

view this post on Zulip Grahame Grieve (Jan 18 2022 at 12:54):

in general, it would be http://hl7.org/fhir/StructureDefinition/Patient|4.3.0 but I don't know how to set that up in this particular case, and I adopted an internal protocol to reference them as http://hl7.org/fhir/4.3/StructureDefinition/Patient - but I don't think that surfaces to anyone else anywhere?

view this post on Zulip Lee Surprenant (Jan 19 2022 at 16:59):

but I don't think that surfaces to anyone else anywhere?

in the page I linked above (https://build.fhir.org/branches/R4B/versioning.html#mp-version) that "http://hl7.org/fhir/4.0/StructureDefinition/Patient" format is provided as an example for how to claim conformance to a particular version of the base spec.

in FHIR#35780 I suggested that be updated to the "http://hl7.org/fhir/StructureDefinition/Patient|4.0.1" format instead

I'm afraid I don't know what you mean by "I don't know how to set that up in this particular case"


Last updated: Apr 12 2022 at 19:14 UTC