FHIR Comparison to SCIM

introduction =

SCIM is a user management API - see http://www.simplecloud.info/

points of comparison with FHIR

• general API approach
  • same approach, but without history operations
  • similar rules around errors, and a structured error content, though the details are a little different (nothing really important?)
  • no history (no use case for it)
  • there's a build transaction (still to be investigated)
  • there's a conformance statement equivalent

• PATCH is supported by SCIM
  • functionality still to be supported

• contents of the resource
  • the API is more bound to the contents of the resources
  • there are only 2 resources
  • PUT rules require extensive processing of the resource content
  • FHIR has avoided this - as a platform rule, but it might be required for specific implementations
  • PUT = merge, and there are merge rules, though they are not effective in collections (which is where the idea would be useful)
  • server is required to canonicalise some fields

• search
  • uses a filter instead of named parameters
  • processing this is more complicated but it's more capable
  • pagination is predictable (not fixed by FHIR spec)
  • doesn't allow for the server to id searches for paginations

• resource descriptions
  • extensibility model different - would only work for json
  • resources are shallower (2 levels max, some FHIR resources are unlimited)