Re: [MMUSIC] Some more thoughts on the simulcast discussion

[Paul Kyzivat <[e-mail]>]

On 10/15/15 5:00 PM, Harald Alvestrand wrote:
> On 10/15/2015 10:46 PM, Adam Roach wrote:
>> On 10/14/15 15:28, Harald Alvestrand wrote:
>>> On 10/14/2015 07:00 PM, Paul Kyzivat wrote:
>>>> On 10/14/15 3:10 AM, Harald Alvestrand wrote:
>>>> What seems clear is that SDP doesn't get the granularity right about
>>>> the stuff it can reference. So people end up using things that aren't
>>>> quite right, and adjusting their usage to align with what it is
>>>> possible to describe.
>>> We are worrying about the same set of issues.
>>>
>>
>> This is actually the exact problem that RID seeks to solve: by
>> defining a new construct rather than trying to leverage existing ones
>> (which, as you observe, end up being ill-fitting in many cases), we
>> are able to point quite precisely at the *thing* of interest, instead
>> of obliquely waving at some badly-matched proxy for that thing.
>
> That is the good thing about RID: That it is able to point precisely at
> one specific media stream.

I think we should carefully identify what we are talking about within 
the terminology of the taxonomy document.

> [distraction: can two media streams have the same RID but different
> SSRCs? What does that mean if it's legal? Where do we enforce the
> restriction if it's illegal?]

That is an important thing to resolve.

> That was also the good thing about a=SSRC (modulo the collision problem).
>
> Now we "only" have to define:
>
> - What exactly we can state about the thing we're pointing at
> - What language(s) we use to make those statements
> - How those statements are resolved against all the *other* tools inside
> SDP that can be used to make statements about things that might also
> happen to refer to the thing we're pointing at.

This last point is important. (And hard.)

> It's possible that we can define RID in such a way that the answers are
> simple enough that they can be written down and agreed upon quickly.
>
> But I can't see us succeeding without at least considering the issues.

Yup.

	Thanks,
	Paul