Re: [rtcweb] Agenda time request for draft-marjou-rtcweb-audio-codecs-for-interop-01

[Michael Procter <[e-mail]>]

On 13 March 2013 17:49, Adam Roach <[e-mail]> wrote:
> On 3/13/13 11:42, Olle E. Johansson wrote:
>>
>> I don't think everyone gets the that selecting a small subset as mandatory
>> for WebRTC does NOT exclude all other codecs. Feel free to add everything
>> you got. Have fun. Let the best mutual codec with the best mutual properties
>> win the session.
>
> Olle hit the nail squarely on the head. This is exactly the purpose of
> offer/answer negotiation: to find the best intersection of endpoint
> capabilities for the desired communication session. I'm bewildered by the
> fairly consistent misinterpretation that having an MTI codec somehow implies
> that every other codec becomes "MNTI".
>

This seems to be the key point - we aren't dividing codecs into two
buckets, one labelled MUST implement and one labelled MUST NOT
implement.  Codecs outside the MTI set can still be implemented and
will be implemented if it makes sense to the implementer.

Conversely, even if you get an RFC with dozens of "MUST implement
<favourite codec>" in it, there won't be any protocol police to
enforce it if implementers decide it doesn't make sense (economic or
otherwise) for them to support those codecs.  There are plenty of
MUSTs in other RFCs that are routinely ignored, after all.

Michael