Okay, so since this thread is continuing, I will repeat here what I
said in the other thread Eran started. (And will be extra careful with
my finicky HTML editor).
1. There are possible use cases for an #include type of behavior
(webfinger?), but we have so far been reluctant to embrace it in the
spec because of the complexity it entails. One approach would be to
not have it now and allow applications to implement this through
extensions and prove its value.
2. There are use cases for a #replace type of behavior but they appear
at this point to be OpenID-specific and may be better addressed by
extensions.
So parsimony and conservatism in embracing changes at this late
juncture would probably both say: Abandon the concept of linked XRDs
altogether for XRD 2.0
--
--Breno
+1 (650) 214-1007 desk
+1 (408) 212-0135 (Grand Central)
MTV-41-3 : 383-A
PST (GMT-8) / PDT(GMT-7)
