Re: [rtcweb] Video codec selection - way forward

[cowwoc <[e-mail]>]

On 13/11/2013 11:11 PM, cb.list6 wrote:
>
> Why no SHOULD implement vp8 and h248? SHOULD means you will do it 
> unless you have a real good reason.
>

SHOULD doesn't carry any accountability. Anyone will tell you they have 
a "real good reason" to avoid implementing one codec or another. We've 
heard plenty of excuses on this list of why one party or another 
believes it is legally risky to implement VP8 or H.264, respectively. So 
in that sense, it is meaningless.

> MUST is too hard for this WG.  Many implementations have a really good 
> reason to not do vp8 OR h248.
>
> Saying that all webrtc MUST do one or the other or both is disingenuous.
>
> SHOULD for both is as good as we are going to get with this 
> complicated IPR environment.  Using MUST is simply not going to work 
> and we have 10,000 on this mailer to back that up.
>

What you say is true, but if we accept this (and go with SHOULD) we are 
essentially ending up with "no MTI" which leads to the dark side 
(transcoding or dropped calls).

Rock --> WebRTC <-- Hard place :)

Gili