Re: [scim] [Technical Errata Reported] RFC7643 (6001)

[Phil Hunt <[e-mail]>]

Shelley,

I agree with Leif’s suggestion that this is the kind of stuff best deli with in a SCIMbis group.

SCIM is badly in need of a certification test that captures common implementation and normalizes behaviours and best practices. In and of itself would be a good feed back to any SCIMbis work.  I have been trying to lobby for this, but without a major sponsor and host organization, I fear this will not happen.

Phil Hunt
[e-mail]



> On Mar 11, 2020, at 6:48 AM, Shelley <[e-mail]> wrote:
> 
> The more I think about this, the more I am leaning towards the problem here being in Section 2.3.7, not 8.7.1.
> 
> As noted here [1], I have some general concerns with "caseExact," namely that there should be some separation of concerns between attribute characteristics and filtering capabilities. The fact that Section 2.3.7 requires all "reference" types to be "case exact" [2] seems overly-limiting.
> 
> URNs, for instance, are not entirely case-sensitive; they define a lexical equivalence, in which only some parts of the URN are case-sensitive [3]. For example, the following two URNs are lexically equivalent:
> urn:ietf:params:scim:schemas:core:2.0:User
> URN:IETF:params:scim:schemas:core:2.0:User
> 
> Given this, I would suggest that a filter for a URN attribute should match on either of the above. However, with the requirements outlined in Section 2.3.7, a client would need to provide the case-exact equivalent in order to match.
> 
> Therefore, perhaps Section 2.3.7's requirement that "A reference is case exact" could/should be removed (in which case my initially proposed change in Section 8.7.1 is optional)?
> 
> [1] https://mailarchive.ietf.org/arch/msg/scim/7ElDxpsgHZGRmQrXOTIR1jwuTts/ <https://mailarchive.ietf.org/arch/msg/scim/7ElDxpsgHZGRmQrXOTIR1jwuTts/>
> [2] https://tools.ietf.org/html/rfc7643#section-2.3.7 <https://tools.ietf.org/html/rfc7643#section-2.3.7>
> [3] https://tools.ietf.org/html/rfc8141#section-3 <https://tools.ietf.org/html/rfc8141#section-3>
> On Wed, Mar 11, 2020 at 4:05 AM Leif Johansson <[e-mail] <mailto:[e-mail]>> wrote:
> 
> 
> On 2020-03-10 21:00, Shelley wrote:
> >     it is allowed that the schema may be set to caseExact false 
> > 
> > 
> > Is Section 2.3.7 not normative? If not, perhaps that section should not prescribe a caseExact characteristic for references at all then.
> > 
> 
> I think Phil meant that the /examples/ are not normative.
> 
> I'm also supportive of going either way on this. There has been talk of a
> 'bis' process for SCIM and if that happens this would certainly get included.
> 
>         Cheers Leif