Re: [Geopriv] Location in SIP and "retransmission-allowed"

From: Henning Schulzrinne ^lt;>
Date: Wed May 02 2007 - 09:08:20 EDT


You keep ignoring the distinction between non-identifiable location
information and a PIDF-LO (personally identified) location object.
Until this distinction is made, including its impact on privacy, we
have no rational basis for discussion.

Please also see my next message about how SIP routing works and that
the proposal you make has no basis in SIP routing or in accepted
privacy standards. This discussion is indeed about SIP, not some
other hypothetical protocols.


On May 2, 2007, at 8:46 AM, Richard Barnes wrote:

> Henning,
> I've addressed your specific comment below, but I think we're
> getting down in the weeds, and I'd like to step back a little.
> The point of GEOPRIV is to allow rule makers to clearly express
> their preferences for handling of LI and LOs, in a manner that is
> independent of the application that is using an LI/LO. With that
> in mind, the options before us are:
> 1. Specify that retransmission-allowed=no has the meaning given in
> RFC 4119, with the exception that use of the enclosed LI for a LoST
> query is also allowed.
> 2. Allow proxies to infer from the inclusion of an LO via a
> Geolocation header that the rule maker consents to use of the LI
> for LoST.
> 3. Provide an explicit indicator of whether a LoST query is allowed
> or not (e.g., binary "routing-query-allowed" or multi-valued
> "allowed-usage").
> With respect to (1): I think this violates the principle of
> application-independence, since SIP is not the only application
> that this would influence. If I send an LO to Google Maps, it
> seems reasonable that they should be requested not to use that
> location for a LoST query.
> With respect to (2): This violates the principle of least surprise;
> Jon Rosenberg might say this is a "DWIM" instead of a "DWIS" approach.
> This leaves us with (3), explicitly indicating that LI may be used
> for LoST. IMO, it would be more future-proof to provide for a list
> of allowed uses (with LoST as the first instance), but the routing-
> query-allowed binary value provides enough information to provide a
> clear indication about LoST.
> --Richard

Geopriv mailing list
Received on Wed, 2 May 2007 09:08:20 -0400

This archive was generated by hypermail 2.1.8 : Wed May 02 2007 - 09:08:44 EDT