sever metadata

[Brian Campbell]

Is there a need for none as a grant_id_supported value? Wouldn’t it be sufficient to have the lack of a grant_id_supported server metadata entry mean that the AS does not support grant ids? I dunno, maybe it’s fine.

Do we really need grant_management_actions_supported? Are there really many cases where only one of the two actions will be supported? The presence of grant_management_endpoint kinda implies at least one (or an erroneous condition of an endpoint that supports no actions). And grant_management_query or grant_management_revokein scopes_supported also implies support of the associated action. Again, I dunno, but grant_management_actions_supported seems like just unnecessary noise.

‌

[Stuart Low]

I think we probably need to decide on whether we are communicating available actions as booleans or in an enumeration. I like enumerations, @Ralph Bragg prefers sets of booleans. @Brian Campbell in your opinion, is it preferred to have few attributes in metadata with “complex” enumerated values or lots of attributes with booleans? @Takahiko Kawasaki @Filip Skokan ?

[Brian Campbell]

I think that first we need to clearly understand what information needs to be conveyed and why. And then decide how best to model that info in JSON.

[Torsten Lodderstedt]

Can we close this ticket? The current revision works as suggested.

[Dima Postnikov]

@Brian Campbell we are planning to close this one tomorrow in the FAPI call. Let us know if there are any objections.

[Dima Postnikov]

Question answered