Stream: committers
Topic: Invalid codes
Grahame Grieve (Jul 21 2016 at 22:23):
GF#10292 - this task is about fixing a bug in the generator that allowed invalid bindings. The following bindings are illegal:
- ModuleMetadata.type
- ActionDefinition.type
- ElementDefinition.type
- Account.status
- Medication.vcategory
- MedicationStatement.category
- Sequence.type
- StructureDefinition,.type
Grahame Grieve (Jul 21 2016 at 22:23):
@Paul Knapp - please fix the accounting one. That should be CodeableConcept
Grahame Grieve (Jul 21 2016 at 22:24):
some of these others are just errors. Disappointing.
Grahame Grieve (Jul 21 2016 at 22:24):
but some represent real challenges to resolve
Bryn Rhodes (Jul 21 2016 at 22:26):
ModuleMetadata and ActionDefinition can actually be removed, no resources are using them anymore (and the type elements in the resources that have module metadata are Codings now). I just have to move the value sets defined there out, that's why they're still in the build. Should I just remove those?
Grahame Grieve (Jul 21 2016 at 22:27):
yes please
Bryn Rhodes (Jul 21 2016 at 22:51):
So I was using those definitions to get the value sets defined, how do I move those out to a shared definition?
Grahame Grieve (Jul 21 2016 at 22:52):
where are they used now?
Bryn Rhodes (Jul 21 2016 at 22:54):
ActivityDefinition, PlanDefinition, Library and Measure.
Grahame Grieve (Jul 21 2016 at 22:57):
and they all refer to it by the canonical url?
Bryn Rhodes (Jul 21 2016 at 22:58):
Yes.
Grahame Grieve (Jul 21 2016 at 22:58):
so you pick one of them and change the reference to valueset-[x] and put the value set in the same folder as the one that you picked
Grahame Grieve (Jul 21 2016 at 22:58):
leave the others as canonical references
Grahame Grieve (Jul 21 2016 at 22:59):
doesn't matter which one
Bryn Rhodes (Jul 21 2016 at 22:59):
Got it.
Brian Postlethwaite (Jul 21 2016 at 23:45):
Account.status should be required binding right?
Brian Postlethwaite (Jul 21 2016 at 23:45):
Account is still a VERY draft resource.
Grahame Grieve (Jul 21 2016 at 23:47):
yes, it should
Bryn Rhodes (Jul 22 2016 at 00:06):
Okay, I have a fix for ModuleMetadata, am I okay to commit?
Grahame Grieve (Jul 22 2016 at 00:12):
yes
Bryn Rhodes (Jul 22 2016 at 00:14):
Working on ActionDefinition now.
Brian Postlethwaite (Jul 22 2016 at 01:10):
Account status has been done and is now built too
Grahame Grieve (Jul 22 2016 at 01:11):
thanks
Bryn Rhodes (Jul 22 2016 at 01:46):
And I assume I'm okay to commit ActionDefinition too?
Grahame Grieve (Jul 22 2016 at 22:22):
yes
Paul Knapp (Jul 24 2016 at 18:03):
Not a coding? - what would be the value of translations? or text?
Lloyd McKenzie (Jul 24 2016 at 19:15):
@Paul Knapp - migration, support by multiple recipients of the same instance. (Presumption in FHIR is that interfaces should not have to be customized for individual receivers, even if that may have been historic convention.
Paul Knapp (Jul 24 2016 at 19:36):
SO if teh valueset is required what is the 'other' valueset one would translate to?
Grahame Grieve (Jul 24 2016 at 20:45):
Lloyd is assuming - based on wide experience - that different systems will be using different source terminologies, and that our experience is that there is much value in passing both the source code and the required code. This is one of the oldest capabilities in v2
Paul Knapp (Jul 24 2016 at 20:46):
SO then it should not be required and be CodeableConcept?
Grahame Grieve (Jul 24 2016 at 20:47):
well, it would usually be one or the other.
Grahame Grieve (Jul 24 2016 at 20:48):
looks it is code and required now, which is fine. I don't recall why I thought it should be CodeableConcept
Paul Knapp (Jul 24 2016 at 21:12):
ok - we can alwasy fixup later
Last updated: Apr 12 2022 at 19:14 UTC
