Stream: committers
Topic: Shareable and Publishable patterns
Bryn Rhodes (Feb 08 2022 at 17:29):
@Lloyd McKenzie , @Grahame Grieve , I've added proposed Shareable and Publishable patterns in a draft PR: https://build.fhir.org/branches/br-32632-shareable-publishable-patterns/shareable.html, https://build.fhir.org/branches/br-32632-shareable-publishable-patterns/publishable.html. I've never added a pattern to the build, so I'm not sure whether I've done it properly, feedback welcome.
Lloyd McKenzie (Feb 08 2022 at 17:59):
I think you should identify that this pattern is intended to only apply to Canonical resources. Also, experimental should be optional - but we should probably add a meaningIfMissing on experimental that if missing, the resource is presumed to be non-experimental.
The definition of "shareable" as a code is tautological. All resources are, by definition, 'shareable'. You need to better qualify the name. (And perhaps change the label, though I realize that's problematic at this stage.)
Publishable calls out a bunch of elements that don't (and won't) exist as discrete elements on ay resource. The patterns should align with the way resources will actually express these elements...
Bryn Rhodes (Feb 08 2022 at 18:01):
ShareableCanonical?
Lloyd McKenzie (Feb 08 2022 at 18:02):
What do you actually mean by "Shareable"? Discoverable? Referenceable?
Bryn Rhodes (Feb 08 2022 at 18:04):
A canonical isn't really shareable without a URL, otherwise you can't identify it across servers, so that's really what it's focused on, just enough information that the resource can be shared and consumers know what it is, what it basically does (or is supposed to do), and who they can talk to with questions about it.
Bryn Rhodes (Feb 08 2022 at 18:04):
We separated it out as a pattern because in the authoring phase, we routinely need to share draft content that doesn't yet have all the metadata that would be expected to be present for a full publishable artifact.
Lloyd McKenzie (Feb 08 2022 at 18:13):
It's shareable. It's just not referencable.
Lloyd McKenzie (Feb 08 2022 at 18:13):
Everything is 'shareable' (can be passed from one to the other).
Lloyd McKenzie (Feb 08 2022 at 18:15):
And the metadata you've include for 'publishable' is pretty much all optional. I think what you're really getting at there is what information might be of interest in a literature-type publication where stuff like editors, contributors, etc. matters.
Bryn Rhodes (Feb 08 2022 at 18:47):
Well, the term Shareable in this sense has been in use since STU3, so I think that's just a matter of making sure we're clear about what we mean by that.
Bryn Rhodes (Feb 08 2022 at 18:49):
And for Publishable, not just literature-type, but a clinical content repository where that sort of information is critical to providing consumers with the justification and tools they need to make use of the content. Who built it, on what grounds, who endorsed it, what evidence supports it, and so on.
Lloyd McKenzie (Feb 08 2022 at 20:05):
Actually, I would say none of that information is needed to make use of the content. However, it might be needed to evaluate the appropriateness of using it. So it's more about "evaluation" than it is about "publication".
Lloyd McKenzie (Feb 08 2022 at 20:06):
Key considerations being: fit for purpose, trustworthy, accurate?
Grahame Grieve (Feb 08 2022 at 20:10):
yes need to stick with the term shareable here, but the intent is that the resources have enough metadata to make sharing them through a public registry reliable
Lloyd McKenzie (Feb 08 2022 at 20:12):
I'm not dead-set on changing the terms. (Though if it's feasible, that would be desirable.) The critical ask is for the definitions to be clearer and more precise.
Bryn Rhodes (Feb 09 2022 at 23:39):
I'm applying this feedback and I don't understand this statement:
Publishable calls out a bunch of elements that don't (and won't) exist as discrete elements on ay resource. The patterns should align with the way resources will actually express these elements...
Can you be more specific about which elements you're referring to?
Lloyd McKenzie (Feb 09 2022 at 23:56):
I guess I lied. I was referring to editor, reviewer and endorser. All three of these seem like things that should be extensions for most if not all metadataresources, but they're explicit in the pattern. (I have a hard time believing that 80% of the systems that will deal with PlanDefinitions and ActivityDefinitions will have all of these things. May be relevant for registries, but not for most use.)
Bryn Rhodes (Feb 10 2022 at 17:25):
Okay, updates applied:
https://build.fhir.org/branches/br-32632-shareable-publishable-patterns/shareable.html
https://build.fhir.org/branches/br-32632-shareable-publishable-patterns/publishable.html
Last updated: Apr 12 2022 at 19:14 UTC
