[Openid-specs-ab] OIDC Discovery and OAuth2 LRDD

Mike Jones Michael.Jones at microsoft.com
Thu Nov 8 18:36:57 UTC 2012


I think we should speak in person because it seems that we may be talking about different things.  Adding the HAL "_links" fields or LRDD fields to our messages sure seems like adding additional and usually unnecessary information to me.

If all you’re talking about doing is adding a JSON Schema description of some of our data structures to the specs, rather than on the wire, I don’t have an objection to that, provided there’s a volunteer to do it.  But if what we’re talking about is adding “_links” or LRDD information on the wire, I’m strongly against that.

                                                            -- Mike

From: Nat Sakimura [mailto:sakimura at gmail.com]
Sent: Thursday, November 08, 2012 10:28 AM
To: Mike Jones
Cc: Richer, Justin P.; openid-specs-ab at lists.openid.net Group
Subject: Re: [Openid-specs-ab] OIDC Discovery and OAuth2 LRDD

Sounds like you are replying to a different thread...

We are not talking about adding metadata to the discovery document. We are talking about how to express the metadata (discovery document). It is not adding new information.

=nat via iPhone

Nov 8, 2012 13:21、Mike Jones <Michael.Jones at microsoft.com<mailto:Michael.Jones at microsoft.com>> wrote:
From a developer perspective, including unnecessary information is nearly always bad.

Your code still has to understand the meaning of the fields it uses.  Adding metadata doesn’t change that.

                                                            -- Mike

From: Nat Sakimura [mailto:sakimura at gmail.com]
Sent: Thursday, November 08, 2012 10:16 AM
To: Richer, Justin P.
Cc: Mike Jones; openid-specs-ab at lists.openid.net<mailto:openid-specs-ab at lists.openid.net> Group
Subject: Re: [Openid-specs-ab] OIDC Discovery and OAuth2 LRDD

From a developer perspective, a uniform interface is always good,  because I can reuse my codes, probably just use libraries so I do not have write much code, etc.

Nat
On Fri, Nov 9, 2012 at 2:58 AM, Richer, Justin P. <jricher at mitre.org<mailto:jricher at mitre.org>> wrote:
The wouldn't a better approach be to take the constructs of LRDD and move them into the JSON world? Maybe using the HAL linking format that Nat's brought up.

What bothers me is a bespoke solution for a generic problem in OIDC.

 -- Justin

On Nov 8, 2012, at 12:37 PM, Mike Jones wrote:

Part of what makes JSON more popular and successful than XML is that there *isn't* any usually-unnecessary metadata or introspection facilities built into the format.  In my opinion, trying to superimpose this structure on our use of JSON after the fact is both unnecessary and counter to what developers want.

They're voting with their feet and we want their votes.

-- Mike
________________________________
From: Nat Sakimura
Sent: 11/8/2012 10:12 AM
To: Richer, Justin P.
Cc: openid-specs-ab at lists.openid.net<mailto:openid-specs-ab at lists.openid.net> Group
Subject: Re: [Openid-specs-ab] OIDC Discovery and OAuth2 LRDD
P.S. You can see how I was feeling from my blog post

http://nat.sakimura.org/2012/09/16/uri-template-in-openid-connect-provider-configuration-response/

It predates http://datatracker.ietf.org/doc/draft-wmills-oauth-lrdd/

Nat
On Thu, Nov 8, 2012 at 10:58 PM, Nat Sakimura <sakimura at gmail.com<mailto:sakimura at gmail.com>> wrote:
It is a bit late in the game, but I do agree being able to express them in the link based structure (not LRDD though, it needs to be JSON) is nice.

HAL or Hyper-meta schema would be good.

Note: none of them are RFC however, so we need to do something in that respect.

The only reason that it is a flat thing is that there were a strong desire to do very simple thing at the beginning. Maybe OAuth discovery document is simple enough that a flat schema makes sense, but OIDC configuration is complex enough that we may want to consider an alternate format.

Having said that, there is a political issues as well. It is soooo late in the documents life cycle and as we do not want to give the community impression that we are still unstable, whether it is worth pursuing should be evaluated carefully.

Best,

Nat

On Thu, Nov 8, 2012 at 12:43 AM, Richer, Justin P. <jricher at mitre.org<mailto:jricher at mitre.org>> wrote:
One of my longstanding complaints about OIDC Discovery is that while it tries to follow a generalizable process to find the issuer, the document that defines the server configuration is a completely bespoke JSON structure. I hadn't seen this document before, but there was recently an admittedly-incomplete attempt by William Mills to put together a spec to define LRDD based discovery for OAuth2 endpoints and configuration parameters.

http://datatracker.ietf.org/doc/draft-wmills-oauth-lrdd/

Shouldn't we be using some kind of host link-based configuration format like this instead of a new JSON document? Shouldn't we be trying to engage the larger service discovery community as opposed to just pasting something in for OIDC alone?

 -- Justin
_______________________________________________
Openid-specs-ab mailing list
Openid-specs-ab at lists.openid.net<mailto:Openid-specs-ab at lists.openid.net>
http://lists.openid.net/mailman/listinfo/openid-specs-ab



--
Nat Sakimura (=nat)
Chairman, OpenID Foundation
http://nat.sakimura.org/
@_nat_en




--
Nat Sakimura (=nat)
Chairman, OpenID Foundation
http://nat.sakimura.org/
@_nat_en





--
Nat Sakimura (=nat)
Chairman, OpenID Foundation
http://nat.sakimura.org/
@_nat_en

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-ab/attachments/20121108/0de1874b/attachment-0001.html>


More information about the Openid-specs-ab mailing list