[Ecrit] Expert review of draft-patel-ecrit-sos-parameter-03.txt

[Gonzalo Camarillo <Gonzalo.Camarillo at ericsson.com>]

Folks,

I have been asked to perform an expert review of the following draft:

http://tools.ietf.org/id/draft-patel-ecrit-sos-parameter-03.txt

The approach taken by the draft seems OK in general. I have a few 
comments though:

The requirement in Section 3 is too specific because it already assumes 
that the solution will be an indication in the SIP header fields. The 
requirement does not need to make that assumption. I would remove "by 
providing an appropriate indication in the SIP header fields".

In Section 5, the reference to RFC 2234 should be replaced with one to 
RFC 5234.

Also in Section 5, the formal syntax should be rewritten so that it is 
compatible with the ABNF in RFC 3261. RFC 3261 already defines 
uri-parameter as follows:

  uri-parameter   =  transport-param / user-param / method-param
                     / ttl-param / maddr-param / lr-param / other-param

  other-param       =  pname [ "=" pvalue ]
  pname             =  1*paramchar
  pvalue            =  1*paramchar

This document should simply define a new value for pname.

The document does not talk about backwards compatibility. What happens 
if the registrar does not understand the 'sos' parameter? Will it do the 
right thing? Will the UAC detect the failure? Is there a need to define 
an option tag?... the document should address these points.

Cheers,

Gonzalo