[Openid-specs-ab] Spec call notes 4-Feb-13
Mike Jones
Michael.Jones at microsoft.com
Tue Feb 5 02:06:44 UTC 2013
Spec call notes 4-Feb-13
Mike Jones
Edmund Jay
Brian Campbell
Pamela Dingle
John Bradley
Tim Bray
Nat Sakimura
Naveen Agarwal
Breno de Medeiros
Agenda:
New Open Issues
MTI for OpenID Request Object
JOSE poll about whether headers must be understood
New Open Issues:
We went through most of the new open issues before starting the MTI discussion with Breno and Naveen
See the issues themselves for resolutions
We closed many of the registration issues
We left open whether to switch registration to a JSON request format, pending further discussion
We'll probably discuss this on Thursday's call
MTI for OpenID Request Object:
Enables, signed requests, requests for individual claims
Breno suggested that another possible processing rule is to ignore everything outside the request object
Breno suggested that claims requests might be separated from the request object
Breno suggested that max_age could be a parameter, as should other commonly used request parameters
Breno wants the request object to essentially be a JSON serialization of the request parameters
Breno asserted that the claims request need not be signed
Breno said that signing the request object provides security to the OP
whereas he said that the claims reflect security/privacy policies of the RP
Breno wants to hear form others such as Salesforce about whether signed requests should be part of the MTI
We agreed to make max_age and preferred_locales top-level parameters
We agreed that it would be valuable to make the ability to request things independent of whether the request is signed
Mike asked whether he should create a distinct "claims" parameter separate from the "request" parameter
A discussion ensued about what encodings should be used for the claims
Breno suggested %-encoding the JSON rather than base64url encoding it
Base64url encoding is only needed for signature validation
He said that %-encoding is fine for the claims request
Whereas it would just be JSON in the request object
We would use UTF-8 %-encoding for the JSON
This would likely help us make progress on MTI consensus
It makes things more orthogonal, so each request parameter can be considered (mostly) independently
We agreed that doing these changes would help us better understand the choices and move towards consensus
JOSE poll about whether headers must be understood
Karen O'Donoghue sent the message "[jose] POLL(s): header criticality" this morning
This will help close an issue that needs to be closed before working group last call
Outcome B (and possibly C) to the third question would break every JWT and JOSE structure
John and Mike will compose a note to the connect working group about this
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-ab/attachments/20130205/443ac770/attachment.html>
More information about the Openid-specs-ab
mailing list