Thank you all guys for discussing about my rudiment question. I know "openid." is only a namespace to avoid collision between the context , I know openid.sign is calculated for the signature base string without "openid." namespace string, and I know the world including me is quite happy with current implementations.
But the asymmetry looks quite strange to me and just wanted to know the good reason for that during developing experimental implementation to extend current OpenID 2.0 protocol. If there MUST be no parameter name collision, ns parameter in KVF response can be optional. Dropping namespace string can save the payload , but it doesn't matter compared to the AuthReq/Res string length in URL redirect responses. We may drop "openid." namespace even in URL only if parties can ensure there is no parameter name contention. All inconsistent to me. But it's OK . No problem. I'm quite a good person to follow the regulations. I'll focus on more productive area. Thank you very much anyway. ---- hdknr _______________________________________________ specs mailing list [email protected] http://lists.openid.net/mailman/listinfo/openid-specs
