Stream: implementers
Topic: Provenance changes
Grahame Grieve (Jan 01 2017 at 10:35):
@John Moehrke I do not understand the changes around Provenance.agent.relatedAgentType - who is the related agent? the R2 model is simple - a list of related agents and types. How does what is there now work?
Grahame Grieve (Jan 01 2017 at 10:35):
and what happened to userId
Grahame Grieve (Jan 01 2017 at 10:53):
nor do I understand why reason and activity changed from CodeableConcept to Coding.
John Moehrke (Jan 01 2017 at 15:50):
The high level answer is that people submitted CR, we discussed them, and voted...
John Moehrke (Jan 01 2017 at 15:53):
Specifics. Provenance.agent.whoUri is where userId goes. This [x] use should likely be done in AuditEvent to help with symmetry.
John Moehrke (Jan 01 2017 at 15:56):
Unless you have a userId that is a proper Identifier, then it goes into Provenance.actor.whoReference.identifier
John Moehrke (Jan 01 2017 at 16:00):
I can't help on Provenance.agent.relatedAgentype. This change was argued, and I lost. I didn't understand the old or the new use-case; especially from the 80% perspective. But when votes are moved and pass, that doesn't matter... until someone else comes up with a new CR, and persuades the committee.
John Moehrke (Jan 01 2017 at 16:02):
on CodeableConcept vs Coding... Can you give an example where this is critical? The desire was to force codes to be used, and the variance one often needs for rapidly expanding code sets didn't seem to apply to a constrained use-case of Provenance. So we simplified.
John Moehrke (Jan 01 2017 at 16:03):
Repeating... changes because someone submitted CR, discussion happened, and vote taken. I welcome more participation.
Grahame Grieve (Jan 01 2017 at 17:35):
I think the CodeableConcept vs Coding is a mistake, personally. I understand that best practice would be to code, but what code sets are there that can be used? If I'm populating a Provenance from an AuditEvent (one use case), I have to use purposeOfEvent, which changed from Coding to CodeableConcept - that's weird, then
Grahame Grieve (Jan 01 2017 at 17:36):
CodeableConcept is not only for 'rapidly expanding code sets', but also for where systems may used different code systems - and that seems particularly applicable here where there is no single well-established code system
Grahame Grieve (Jan 01 2017 at 17:37):
userId - ok, I understand that
John Moehrke (Jan 01 2017 at 17:37):
we welcome the informed advice of the Director. File a CR.
Grahame Grieve (Jan 01 2017 at 17:41):
Grahame Grieve (Jan 01 2017 at 17:50):
and GF#12502
Grahame Grieve (Jan 01 2017 at 17:59):
as for Provenance.entity mapping between R2 and R3 - I thought there was an open task about Provenance.entity.reference needing to be a non-reference? but I can't find it?
John Moehrke (Jan 01 2017 at 18:13):
yes, that was in chat, but didn't get a CR written on it... go ahead and put one in.
Grahame Grieve (Jan 01 2017 at 18:23):
Last updated: Apr 12 2022 at 19:14 UTC
