Re: [rtcweb] Quick comments on draft-roach-rtcweb-glareless-add-00

[Paul Kyzivat <[e-mail]>]

On 5/8/13 5:18 PM, Adam Roach wrote:
> On 5/8/13 15:07, Paul Kyzivat wrote:
>> The persistent answerer must send a solicitation that describes the
>> offer it wants the persistent offerer to make on its behalf. You don't
>> explain how this would be encoded, but in general it must be like a
>> delta on the last offer it received.
>
> Sorry, I thought I'd spelled this out relatively clearly:
>
>     So, for example, if the
>     "persistent answerer" wishes to add a new audio-visual element, it
>     sends a solicitation to the "persistent offerer" indicating that it
>     requires one new audio m-line section and one new video m-line
>     section.
>
>
> What I was trying to say above is that the only information you would
> need to convey is literally one, two, or three <mediatype,ordinality>
> pairs. You don't say what you're planning to do with them, just that you
> need them.

How is that possibly enough? What codecs and codec parameters/optons? 
What bandwidth? What bundling options? The list goes on.

> I intentionally left syntax out of the document because giving a
> non-normative exemplary syntax seems to encourage infinite bikeshedding
> rather than the high-level discussion that is necessary to form consensus.
>
> So, at the risk of encouraging bikeshedders to complain about dictionary
> key names or XML attributes versus cdata, I'm envisioning something that
> looks roughly like this (using JSON format):
>
> {
>    'newAudioStreams': 1,
>    'newVideoStreams': 1,
>    'newDataChannels': 0
> }
>
> ...or like this (in XML):
>
> <solicitation>
>    <newAudio count="1"/>
>    <newVideo count="1"/>
>    <newData count="0"/>
> </solicitiation>
>
>
> (Given that this isn't necessarily subject to standardization -- at
> least for the WebRTC-only cases --  implementations could really send
> something as minimal as "[1,1,0]" to get the same point across.)
>
>> I remain dubious that this glare situation is a problem that needs to
>> be fixed.
>
> So am I, but Justin seems to be obsessed with it. That's why I'm
> proposing a non-normative recipe that allows people who are
> unnecessarily preoccupied with the issue to indulge their peccadillos.
> The rest of us are free to ignore it.

Well, I agree that if somebody wants to do this on their own, then that 
is their business. If applied in a sufficiently constrained environment 
the amount of info that needs to be conveyed can be reduced, as you 
suggest. And of course the token passing approach requires very little 
to be passed.

But I don't want to standardize any of this.

	Thanks,
	Paul