Re: [rtcweb] MTI Video Codec: a novel proposal

[Harald Alvestrand <[e-mail]>]

On 11/11/2014 01:02 PM, cowwoc wrote:
> On 11/11/2014 3:13 PM, tim panton wrote:
>> My personal preference would be to replace ‘both’ with ‘either’ in
>> the device rule, but I understand
>> the thinking behind the ‘both’ phrasing.
>
> Same here. I would also prefer replacing "both" with "either" in the
> device rule. What is the practical benefit of forcing both codecs on
> non-browsers? If you proceed this way, won't most applications accept
> being declared non-compliant instead of investing the extra work/cost
> of implementing both codecs?
>
> Implementing both codecs only makes sense when you need
> interoperability between applications written by different authors.

Don't forget the case where one author wants to write the same
application in two different environments, and wants to use some
toolkits not written by himself.

If those toolkits don't support the same codecs ... he can't
interoperate with himself.

> The same argument does not apply when both endpoints are written by
> the same author. Furthermore, I would argue that when separate authors
> want their applications to inter-operate they will have the necessary
> incentive to make that happen, with or without the specification
> dictating how it should be done.
>
> Another spin on this would be to say: all libraries/platforms
> (abstraction layers used to write applications) must implement both
> codecs, whereas applications (end-users) can implement as many codecs
> as they'd like.

That would indeed be a possible spin, but a somewhat strange one.

>
> Gili
>
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb


-- 
Surveillance is pervasive. Go Dark.