-
Type:
Change Request
-
Resolution: Persuasive with Modification
-
Priority:
Medium
-
FHIR Core (FHIR)
-
STU3
-
FHIR Infrastructure
-
ValueSet
-
Value set uri
-
-
Grahame Grieve/Ewout Kramer: 14-0-7
-
Clarification
-
Non-substantive
-
STU3
Comment:
http://hl7.org/fhir/2018Sep/datatypes.html#canonical defines "canonical" as "A URI that refers to a canonical URI."
This is prima facie confusing. Does this mean that the subject URI a) is a canonical URI, b) that it contains one, c) that it is-contained by one, d) that it refers to some resource from which the canonical URI can be derived, e) or something else?
The XML representation of canonical-"xs:anyURI"-doesn't help clarify the issue.
"Note that the type canonical is not used for the actual canonical URls that are the target of these references, but for the URLs that refer to them, and may have the version suffix in them."
This seems to rule out (a), but it leaves 'refer' unclear.
"Canonical URLs may also have #fragment references"
Any URI may have a #fragment reference. Does this constrain non-canonical URLs from doing so?
http://hl7.org/fhir/2018Sep/references.html#canonical :
"Many resource types have a defined element "url" which is the canonical URI that always identifies the resource. The canonical is the preferred way to reference a conformance or knowledge resource."
This seems clear enough, though the use of "canonical" as a noun is a little disorienting, and it might be better to specify "resource instance."
However, each of the resources identified here as using 'canonical' has a property "uri" of type "url" with text asserting it's 'canonical.' I.e., the 'canonical' type is not used; rather, the set of expected behavior may be asserted in text.
http://hl7.org/fhir/2018Sep/references.html#canonical
"References to canonical URLs may include a version"
Both URI and Canonical have the same xml definition. Does this imply that non-canonical uris may not include a version?
I believe the source of confusion for me is that the canonical identifier is documented as a type but it's really a property.
1. Rename the property from "uri" to "canonicalUri" to avoid confusion with the commonly understood term "uri." The datatype seems redundant, and it could be discarded (though it's used in ElementDefinition).
2. Clarify the relationship.
Description of property: change from "Canonical identifier for this value set, represented as a URI (globally unique)" to "Globally unique URI for this value set"
Definitions: change from "This URL can be the target of a canonical reference." to "This URL can be the target of a canonicalUri reference in a resource."
http://hl7.org/fhir/2018Sep/datatypes.html#canonical: idealy, remove the type, but if we must,
- is voted on by
-
BALLOT-6056 Negative - Greg Staudenmaier : 2018-Sep-FHIR R4 TERMINOLOGY
- Balloted