Re: [AVTCORE] Fwd: New Version Notification for draft-terriberry-avp-codecs-00.txt

[Stephan Wenger <[e-mail]>]

Hi,
As I said in the meeting, I disagree with Colin that an update of 3551 is
needed.  Rtcweb is producing essentially a suite of profiling/systems
specs, and many other IETF technologies are profiled therein.  I don't see
why AVPF needs to be an exception.
That said, if the group chooses to follow Colin's arguments, should the
resulting (new) profile not have a different name, registration, etc.?
Also, while we are at it, shouldn't we do away with other known and
commonly agreed shortcomings of AVPF?  All pre-reserved payload types come
to mind...
Finally, what about SAVPF?  In the past, we tried to align the two.
Shouldn't this draft not at least mention that it is intended to update
SAVPF as well?
I'm not trying to be difficult here, and I don't want to open a can of
worms larger than necessary.  However, single-use, rtcweb local fixes
ought to be dealt with in rtcweb, and not here.  If these fixes ought to
have general applicability, then, I think, we need to consider thins a bit
more broadly.
Stephan


On 8.7.2012 14:32 , "Timothy B. Terriberry" <[e-mail]> wrote:

>Colin Perkins pointed out that if we do not want to make implementing
>the DVI4 codec a SHOULD in rtcweb, then we need to update the AVP
>profile which rtcweb relies on. I also added, at Colin's suggestion, a
>sentence noting that G.711 may be mandatory in some environments.