Torsten,

I obviously agree that supporting responses with multiple tokens is useful.
My original suggestion (application/credentials 
http://www.ietf.org/mail-archive/web/oauth/current/msg01920.html) also had 
multiple tokens with just one refresh_token.

I now think it would be simpler to understand, parse, specify, and agree on if 
we don't try to split fields between those that are common to all the tokens in 
a response and those that are specific to one token. A JSON array of JSON 
objects -- one JSON object per token -- is significantly simpler. If some 
information is duplicated in two tokens that is a very minor inefficiency (eg a 
few dozen extra bytes in one response).



Marius,
Supporting the swapping (down-scoping) of tokens with extra calls might be 
feasible, but it feels like a more complex and less flexible solution. It has 
more overhead (extra round trips). It would also need extra information to tell 
the client that swapping is possible, and for what scopes/servers/schemes/... 
it can, should or must be done.

> The original access token (the one that came along the refresh token) would 
> be discarded.

This doesn't meet some of the requirements. Tokens for HTTP and HTTPS are 
needed simultaneously, not sequentially. Similarly with tokens for a 'contacts' 
server and a 'calendar' server; or tokens to use with BASIC and TOKEN schemes 
etc.

-- 
James Manger


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

Reply via email to