I don't know that we need user_id in the JWT spec it may be enough to have it 
as a OIDC extension if it is not globally useful.

I agree that the definition of prn should be more specific.
On 2012-11-26, at 3:56 PM, Torsten Lodderstedt <[email protected]> wrote:

> Hi John,
> 
> does it make sense to change the spec as follows:
> 
> - specify the prn claim as being globally unqiue
> - add user_id as scoped by iss claim
> 
> What do you think?
> 
> regards,
> Torsten.
> 
> Am 26.11.2012 19:51, schrieb John Bradley:
>> A user_id is scoped to a iss.
>> 
>> There is some notion that a prn is globally unique, though the JWT spec is 
>> not clear on that.   I think people are thinking of it like a UPN in LDAP/AD.
>> 
>> John B.
>> On 2012-11-26, at 3:46 PM, Torsten Lodderstedt <[email protected]> 
>> wrote:
>> 
>>> Hi John
>>> 
>>> thanks for the explanatian. Just to make sure I got you right. A prn can be 
>>> a user_id. A prn is bound to the scope of an iss.
>>> 
>>> Regards,
>>> Torsten.
>>> 
>>> 
>>> 
>>> John Bradley <[email protected]> schrieb:
>>> JWT is more generic than OIDC.
>>> 
>>> prn and user_id as used by OIDC are similar.   user_id is already in wide 
>>> use with Facebook's signed request.  
>>> We were hoping that Facebook would be more likely to migrate from signed 
>>> request to JWT if the parameter names stayed the same for developers.
>>> 
>>> In the generic case of a JWT the prn may not be a user.   
>>> 
>>> The other discussion that I recall around prn was a notion that they are 
>>> fully qualified and globally unique.
>>> 
>>> We wanted to be clear with user_id that it is scoped to the iss and not 
>>> globally unique.
>>> 
>>> So a prn was seen as a User Principal name and the user_id was seen as a 
>>> persistent non reassignable identifier for the user in the context of the 
>>> iss.
>>> 
>>> John B.
>>> 
>>> 
>>> On 2012-11-24, at 3:47 PM, Torsten Lodderstedt <[email protected]> 
>>> wrote:
>>> 
>>> Hi,
>>> 
>>> I've got a few comments on your draft.
>>> 
>>> I’m wondering why neither acr nor auth_time (which are used in OIDC) made 
>>> their way into this spec?
>>> 
>>> What is the difference between prn and the user_id claim OIDC uses?
>>> 
>>> regards,
>>> Torsten.
>>> 
>>> 
>> 
> 

_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to