Stream: cql
Topic: Guidance for URLs
Corey Sanders (Jan 14 2021 at 16:02):
We have authors creating custom libraries and measures inside our organization. Is there any guidance on how to define the URL for these type of measures? And should the canonical URL always be used in the depends-on relationships? There seems to be some inconsistency in the FHIR examples and connectathon resources. I see some depends-on relationships using a canonical URL such as http://hl7.org/fhir/Library/FHIRHelpers|4.0.1 and then others using a resource and resource id such as Library/library-FHIRHelpers-4.0.1. I know both are technically valid, but I'm looking for best practice recommendation. I feel like maybe I should be making up my own URL prefix (http://company.com/fhir/ or something like that) and using that for all of the related artifact declarations.
JP (Jan 15 2021 at 20:34):
Canonical url is preferred for terminology, conformance, and knowledge resources (ValueSets, Measures, Libraries, PlanDefinitions, etc) as of R4.
https://www.hl7.org/fhir/references.html#canonical
The url prefix is supposed to match whatever is defined as the base url in the the IG which hosts the knowledge content. The cql-translator provides a namespace manager for mapping prefixes to namespaces --there's a 1:1 relationship. The base url (and an IG) is generally supposed to group related artifacts.
How the url is resolved is implementation specific but it's supposed to first search on a known registry of resources, and then fall back to direct url resolution.
Last updated: Apr 12 2022 at 19:14 UTC
