FHIR Chat · Logical Model and differential · conformance

Stream: conformance

Topic: Logical Model and differential


view this post on Zulip Alexander Henket (May 16 2019 at 11:38):

@Michel Rutten / @Grahame Grieve Typically logical model are snapshot only. For the logical models under fhir-cda however there is even a derivation relationship, profile-on-profile if you will.

In V3 that this project is modeling, all datatypes inherit from an abstract datatype ANY. Each datatype then either adds or constrains other attributes/elements. Example:

CS constrains CV, CV constrains CE, CE constrains CD, CD extends ANY.

This means (in my mind) that differential matters. Forge currently disregards the differential and leaves it as-is. Should the differential support in logical models not be beefed up?

view this post on Zulip Grahame Grieve (May 16 2019 at 12:05):

that's a forge question, I think. the answer for the validator/ig-publisher is 'yes. and I think that's already how it is'

view this post on Zulip Alexander Henket (May 16 2019 at 13:34):

Ok so maybe this a Forge thing. I've tested what would happen if I delete the snapshot and let Forge operate solely on the differential. Forge will happily open it. Upon save (with snapshot), the snapshot is created but misses the base properties, and also the differential is gone at that point.

view this post on Zulip Michel Rutten (May 20 2019 at 09:58):

Forge support for Logical Models is based on feedback from the core team.

  • When opening a Logical model in Forge, the element definitions are loaded from the snapshot component, if present, or from the differential component otherwise.
  • When saving a logical model, Forge serializes the list of element definitions to the snapshot component. The differential component is always cleared.

view this post on Zulip Grahame Grieve (May 20 2019 at 10:33):

I don't think that the second is our advice. At least, I think it's not right


Last updated: Apr 12 2022 at 19:14 UTC