Stream: conformance
Topic: Logical Model for EN and AD
Alexander Henket (May 16 2019 at 11:44):
@Michel Rutten / @Grahame Grieve Normally the order for paths in StructureDefinitions is what determines order in the instances. For V3 logical models this does not hold.
- There is no determined order for XML attributes
- There is no determined order for EN particles of type ENXP and no order for AD particles of type ADXP
I see an indicator in the EN/AD datatypes http://hl7.org/fhir/StructureDefinition/structuredefinition-xml-no-order that aims to support this, partly. The thing is that there is at least 'some' order:
- Attributes come before child elements
- EN.validTime will always be trailing all other child elements
- AD.useablePeriod will always be trailing all other child elements
I'm wondering if the current expression is precise enough?
Grahame Grieve (May 16 2019 at 12:03):
attribute order is irrelevant.
Grahame Grieve (May 16 2019 at 12:04):
the current expression does not capture the constraint that validTime/useablePeriod comes last. I did not think that I needed to capture that when I did work the first time around, but if we're going to do real validation, I guess we do
Grahame Grieve (May 16 2019 at 12:04):
are there other similar cases anywhere?
Alexander Henket (May 16 2019 at 13:41):
Derived datatypes from EN: ON, PN, TN
Sort of similar: IVL allows combinations of low + (width|high), high, width + high, center + width. StrucDef will probably need constraints for that.
ED has a text() node somewhere, but reference and thumbnail have to be in that order
I guess that's it
Grahame Grieve (May 16 2019 at 21:19):
ok I'll think about those. maybe all of them done through FHIRPath?
Last updated: Apr 12 2022 at 19:14 UTC
