Re: [Geopriv] AD Review: draft-ietf-geopriv-held-identity-extensions-04

["Thomson, Martin" <[e-mail]>]

On 2010-09-29 at 04:26:30, Robert Sparks wrote:
> Summary: This draft is ready for IETF LC - I'll be requesting that LC
> shortly.
> 
> Please consider the following as early LC comments:
> 
> The last sentence of (1.1 Applications) is a bare assertion.
> Can you add text or point to more discussion?

Good point.  Would this help?

Section 1.1 (last paragraph), ADD:
  For instance, a policy rules document could be used to express the agreed policy.  Formal policy documents, such as [RFC4745], can be applied in an automated fashion by a LIS.

> The first sentence of the third paragraph of section 2.1
> (Starting "The identifiers described") is very hard to parse.
> Could it be simplified to "The identifiers described in this
> document MUST only be used as inputs for location determination."?
> Can you call out some example applications where it is not ok
> to use these identifiers (helping to motivate the MUST).

That text sounds OK.  Actually, since this is just a (poor) paraphrase of RFC 5687, maybe we could instead lift that text:
 
      The LIS MUST be able to use the identifier [...] for location determination. 

Would this plagiarism work for you?

> In section 3.5's discussion about ensuring a URI refers to a single
> device, please consider pointing to GRUU.

It was always the intent to use GRUU :)

Section 3.5, OLD:
   ...  For example, a SIP address of record URI can
   correspond to a service subscription rather than a single Device.
NEW:
   ...  For example, a SIP address of record URI can
   correspond to a service subscription rather than a single Device;
   a Globally Routable User Agent URI (GRUU) [RFC5627] might be more
   appropriate.

Cheers,
Martin