Re: [scim] Common Attribute "schemas" Characteristics

[Shelley <[e-mail]>]

Regarding the "*returned*" characteristic for the "*schemas*" attribute...
RFC 7644 includes examples with conflicting behavior for the "schemas"
returned characteristic.

In section 3.4.2 [1], the example query which specifies "attributes" does *not
*return the "schemas" by default:

GET /Users*?attributes=userName*
{
 "schemas":["urn:ietf:params:scim:api:messages:2.0:ListResponse"],
 "totalResults":2,
 "Resources":[







*   {     "id":"2819c223-7f76-453a-919d-413861904646",
 "userName":"bjensen"   },   {
 "id":"c75ad752-64ae-4823-840d-ffa80929976c",     "userName":"jsmith"   }*
 ]
}

However, in section 3.9 [2], "schemas" *is *returned by default in the
response, despite the "attributes" parameter not specifying this attribute:

GET /Users/2819c223-7f76-453a-919d-413861904646*?attributes=userName*
{
 "schemas":["urn:ietf:params:scim:schemas:core:2.0:User"],

* "id":"2819c223-7f76-453a-919d-413861904646",* "userName":"bjensen"
}

RFC 7643 generally indicates throughout that "schemas" MUST be specified;
however, it is not clear as to whether the requirement is referring to
requests and/or responses. Unlike the "id" attribute, which clearly
specifies that it is expected to have a returned characteristic of
"always", the other common attributes don't clearly specify this
characteristic. Is "default" the correct attribute, such that the example
in RFC 7644 Section 3.9 is not correct?

[1] https://tools.ietf.org/html/rfc7644#section-3.4.2
[2] https://tools.ietf.org/html/rfc7644#section-3.9

On Fri, Mar 13, 2020 at 12:10 PM Phillip Hunt <[e-mail]>
wrote:

> Shelley
>
> In practical/pragmatic terms...case insensitive.
>
> I treat as case insensitive since it is unlikely and unwise if two
> separate schemas only differentiated by case.
>
> I would also expect iana process to reject “overlapping” registrations
> differentiated only by case.
>
> Phil
>
> On Mar 13, 2020, at 9:16 AM, Shelley <[e-mail]> wrote:
>
> 
> What are the characteristics for the "schemas" [1] attribute?
>
> Here is my attempt at defining these characteristics using the Schema
> definitions:
>
>    - "type" is "reference"
>    - "referenceTypes" is ["uri"]
>    - "required" is "true"
>    - "multiValued" is "true"
>    - "uniqueness" is "none"
>    - "caseExact" as "true" (since this is a "reference" type)
>    - "mutability" of "immutable" (although none of the mutability values
>    seems like a perfect fit)
>    - "returned" characteristic of "default"
>
> Although "schemas" and the "Common Attributes" don't define their own
> schemas, it would be nice to have all of these attributes' characteristics
> clearly defined in the spec using the Schema definition to help provide a
> clear/common definition, particularly, since these characteristics are not
> intended to be modified/defined by SPs..
>
> In particular, in my SCIM implementation, I have been considering whether
> to evaluate "schemas" in resource representations* case-sensitively*
> (exact case as defined in the ResourceType),* case-insensitively* (any
> case allowed), or using *lexical equivalence* (e.g. for URNs,
> case-insensitive schema and NID, case-sensitive NSS, and some components
> ignored). The RFC doesn't seem to clearly prescribe this, but based on the
> fact that the implied type is "reference" which (debatably [2]) has
> "caseExact" as "true" and the fact that the attribute "MUST only contain
> values defined as "schema" and "schemaExtensions"", I'm under the
> assumption that this attribute is *case-sensitive*.
>
> Please confirm. Thanks!
>
> [1] https://tools.ietf.org/html/rfc7643#section-3
> [2]
> https://mailarchive.ietf.org/arch/msg/scim/05_K_y-V26EOfN2F7fuSO3DXoLw/
> _______________________________________________
> scim mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/scim
>
>