This section makes sense as it does setup for the identification and 
authentication of the client, so I do believe that it makes sense in the 
document.  Thomas undertook the task to revive the Client Credentials section 
from previous draft that will include normative text, which when asked the room 
showed interest, I believe he indicated 3 weeks to do this.

I assume you are objecting to not using just HTTP authentication? There are 
scenarios where we need SAML and other technology so only allowing HTTP 
authentication is too restrictive and to allow the profiling for 
interoperability of these technologies we need an extension point here.


From: [email protected] [mailto:[email protected]] On Behalf Of Eran 
Hammer-Lahav
Sent: Friday, April 01, 2011 2:14 AM
To: OAuth WG
Subject: [OAUTH-WG] Reques to drop section 3

There was (still is) a long heated debate at the WG meeting today about client 
authentication and the dropped client assertion credentials section. I want to 
repeat my past view (and this time post it as an open issue), that this entire 
section makes no sense in this document. OAuth should not be defining hackish 
HTTP authentication schemes, especially ones not using the RFC2617 framework.

Someone can easily register the client_password parameter as an extension (it's 
a nasty hack but I won't stand in its way), as well as any other poorly design 
client authentication scheme using form-encoded parameters to authentication an 
HTTP request...

So - I want to see section 3 turned into a brief discussion about client 
authentication which gives HTTP Basic auth as an example and nothing else. 
Client authentication is already 95% out of scope.

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

Reply via email to