On 5 Nov 2006, at 19:44, Gorry Fairhurst wrote:
>
> Colin, I have just a few queries (mainly wording) about
your I-D.
>
> Best wishes,
>
> Gorry
>
> ---
>
> In Section 4.2, you don't define or reference SIP when
first used:
> " for example by performing a SIP
re-invite..."
> - Perhaps say: "Session Initiation Protocol, SIP
[24]"
Will fix.
> In Section 4.2 (towards the end) the I-D says:
> " If there is
> overlap between RTCP report packets and DCCP
acknowledgements, an
> application should use either RTCP feedback or DCCP
> acknowledgements,
> but not both "
> - Should this be phrased as an RFC2119 format
"SHOULD"?
> - or is this really just a user-choice?
This is a user choice, I think. The right answer would
likely depend
on the scope of the RTP session. I propose to make no
change.
> In Section 4.4:
> " The use of a unicast DCCP connection does not
imply
> that the RTP session will have only two
participants, and RTP end
> systems must assume that multiple synchronisation
sources may be
> observed when using RTP over DCCP."
> - Should this be phrased as an RFC2119 format
"MUST"?
Possibly a SHOULD, since there may be out of band reasons
why the
session is known to have only two participants.
> In Section 5.1:
> " The "DCCP" protocol identifier MUST
NOT be
> used to signal RTP sessions running over
DCCP..."
> - Not sure the meaning of this is clear, does it refer
to the use
> of the
> string "DCCP" on its own?
Yes. It's the same as plain "UDP" is not used for
RTP sessions.
Colin
|