Re: [Sip] RE: [Geopriv] Does Location-Conveyance have an HTTP dereferencemechanism defined

From: Ted Hardie ^lt;hardie@qualcomm.com>
Date: Wed Aug 02 2006 - 14:52:06 EDT

At 9:24 AM -0400 8/2/06, Andrew Newton wrote:
>Henning Schulzrinne wrote:
>>I'll also re-iterate and rephrase another point made by somebody else before: Restricting URL schemes makes the system non-upgradeable. If a proxy supports "RFC-conveyance" that restricts URLs to a particular scheme list, it will have to reject as flawed a request that contains a newly-defined protocol it doesn't yet know about. Thus, reliably introducing a new protocol would have to wait until every location-aware proxy in the world has been upgraded, which I hope we agree is untenable. The right model is that a proxy that doesn't implement or doesn't want to support (by policy) a certain URL scheme is to ignore it. SIP always has strived to make it possible to evolve systems incrementally, so this type of restriction is well outside established SIP design practice.
>
>Henning, these are separable issues. Codification of rejection of unknown URI schemes and maintenance of a list sanctioned schemes are two different things.
>

Henning,

What difference do you see between the proxy behavior when something
is rejected as "flawed" because it contains a newly defined protocol and proxy behavior
when the URL scheme list is open and the proxy encounters one it does not understand?

Ted

_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www1.ietf.org/mailman/listinfo/geopriv
Received on Wed, 2 Aug 2006 11:52:06 -0700

This archive was generated by hypermail 2.1.8 : Wed Aug 02 2006 - 15:19:47 EDT