Re: [OAUTH-WG] For review/discussion: Cedar profile of OAuth Rich Authorization Requests

2024-02-23 Thread Justin Richer
On top of that, since the RAR structure is fundamentally an array, existing open banking stuff could still use their types alongside the cedar type, if they wanted. Combining them would be an interesting exercise for each ecosystem, but you’ll run into that regardless of the vertical. — Justin

Re: [OAUTH-WG] For review/discussion: Cedar profile of OAuth Rich Authorization Requests

2024-02-23 Thread Brian Campbell
Yeah, communication of intent was not supposed to be the purpose of the type parameter value. Although I do (now) see how many of the examples in RAR/RFC9396 kinda read that way. The https://www.rfc-editor.org/rfc/rfc9396.html#section-2-2.2 definition of type tries to convey what it is intended to

Re: [OAUTH-WG] For review/discussion: Cedar profile of OAuth Rich Authorization Requests

2024-02-23 Thread Cecchetti, Sarah
Interesting. We considered using the type parameter, but decided against it. In the examples in the spec, the spirit of type seems to be an indication of the intent of the request (for example "customer_information" or "payment_initiation.") We were concerned about breaking existing open

Re: [OAUTH-WG] For review/discussion: Cedar profile of OAuth Rich Authorization Requests

2024-02-23 Thread Brian Campbell
I'm inferring some intent (apologies if I've got it wrong!) but I think it'd make the most sense for this work to start with defining a RAR type value (something like "https://cedarpolicy.com;) and define that type as having the "policySet" parameter. An updated example figure 1 from the draft

Re: [OAUTH-WG] For review/discussion: Cedar profile of OAuth Rich Authorization Requests

2024-02-22 Thread Justin Richer
Hi Sarah, Thanks for putting that draft together. As one of the authors of RAR, I wanted to chime in. First, I do think that this is a great use of RAR. The whole idea behind RAR was to give people structures that they could use beyond what scopes allow, and tying this to a computable policy