Modularizing Auth 2.0 Discovery

[Drummond Reed]

>Martin Atkins wrote:
>
>Attached is a version of the current spec draft 11 with large chunks of 
>the discovery/normalization stuff hacked out and replaced with generic 
>"see other specifications" text.
>
>I think I may have deleted too much in that we've lost the detailed 
>rules for dealing with XRDS documents, which are probably going to be 
>referenced by most discovery specs and deserve to be in the core spec. 
>See the TODO: markers for more info.
>
>However, in doing this it also occured to me that it would be very 
>useful to have the XRDS-related bits of the XRI resolution spec in a 
>separate document that we could reference, since our new 
>protocol-agnostic core Auth 2.0 spec has no need for the rest of XRI 
>resolution. I wonder if the XRI guys would consider making Chapter 3 of 
>the XRI Resolution 2.0 spec into a separate document that OpenID Auth 
>could reference directly?

The XRI Technical Committee has been working with the OpenID community for
the last year to make sure the XRI Resolution spec is cleanly modular so
that independent pieces of it can be referenced as needed by other specs.
Yadis already does that. Given the extra overhead each spec carries at a SDO
like OASIS, it would be much easier for OpenID Discovery to reference just
the parts of XRI Resolution that it needs. 

But I suggest we close on the best approach for OpenID Discovery first. I'll
answer a different question you posed about that next.

=Drummond