[Openid-specs-ab] Spec call notes 17-May-12
Mike Jones
Michael.Jones at microsoft.com
Fri May 18 00:25:37 UTC 2012
Spec call notes 17-May-12
Mike Jones
Edmund Jay
John Bradley
Pamela Dingle
Nat Sakimura (joined near the end of the call)
Agenda:
Open Issues
Mailing list items
Editing
JWE KDF Parameters
Interop
Discovery
Open Issues:
#584 Messages - Username claim
An alternative would be a local_user_handle claim that the IdP could optionally release
We decided to defer a decision on this one pending more discussion
#587 Registration - 2.1 Should mention about OAuth Bearer Authz Scheme
John pointed out that Bearer access only applies to the client_associate type
He is adding that to the ticket
There's a separate issue about whether we want to be able to use JWT assertions to update associations
(rather than the client_secret)
We're leaving this one open until Nat can also participate in the discussion
Mailing list items:
Amanda Anganes' note "redirect_uri matching clarification"
John replied with additional rationale about registered redirect_uri values
Chuck Mortimore's note about whether to require redirect_uri registration
We either need required nonce verification or required redirect_uri registration
Given we no longer have the first, we need the second
John already replied to that effect
Editing:
Nat checked in his changes, including being able to return claims in the ID Token
Edmund updated the related examples
John checked in changes to remove the Check ID endpoint
The corresponding changes to Basic will happen next
Mike should be able to do his edits today or tomorrow
JWE KDF Parameters:
Edmund said that XML ENC uses no datalen values
If that's the case, we can safely not use them too
John believes we need non-constant PartyUInfo and PartyVInfo values when doing key agreement
XML ENC Key Agreement just says that the values need to be used (but are provided as parameters)
John wonders if we should do the same
PartyUInfo includes an identifier and a nonce
Interop:
We have interoperable issuers with paths
We have interoperable token hash implementations now
Mike needs to include new test descriptions that Roland sent him on the OSIS interop wiki
Discovery:
WebFinger/SWD: We should push for servers to accept e-mail addresses without schemes
Particularly since acct: may or may not get approved
Mike may ask the chair whether to have a consensus call about separating acct:
Open Issue: What do we do when there is no HTTP server for a domain?
Applies to most of the hosted situations
Using DNS in some manner is likely the right solution
Such as reusing the MX record, which while "unclean", would work
Alternatives are SRV or TXT records
John will file an issue about this
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-ab/attachments/20120518/7f4061eb/attachment.html>
More information about the Openid-specs-ab
mailing list