Re: [MMUSIC] Updated MSID draft

[Flemming Andreasen <[e-mail]>]

On 5/4/14, 3:26 PM, Harald Alvestrand wrote:
> On 05/03/2014 04:37 AM, Flemming Andreasen wrote:
>> Hi Harald
>>
>> I have taken a closer look at msid-05 and have a number of comments. I
>> will list the major ones and/or what might require list discussion
>> here and send you a bunch of more detailed ones in an annotated
>> version off-line:
>>
>> 1) There are some grammar inconsistencies between the way you use
>> token and (some) of the text you have. In general, it seems the length
>> should be limited and the set of characters you want to allow are less
>> than what token allows for.
>>
>> 2) Offer/Answer procedures are lacking. We need to have an O/A section
>> with subsections as per RFC 3264 (for unicast) to better specify the
>> overall behavior here (Generating the Initial Offer, Generating the
>> Answer, Offerer Processing of the Answer, Modifying the Session)
> Hm. But MSID isn't an offer/answer field. It's an unilateral
> declaration, and refers to no information coming from the other side.
>
> Does it make sense to follow the offer/answer section format when the
> answer doesn't depend on the offer?
It does. Even declarative SDP parameters have procedures that must be 
followed by the offerer and the answerer and we want those called out 
clearly.
>
>> 3) There is some ambiguity as to what constitues an msid value and
>> more specifically when it includes both the MediaStream identifier and
>> the MediaStreamTrack and when (if ?) it doesn't.
> Hm. Must make this clearer. The generic format has MST-identifier
> optional, and the WMF usage has MST-identifier mandatory.
>
The confusion is more around some of the text in the document which is 
not always clear on this (to me at least).


>> 4) Unified-plan is being referenced in a normative manner (but is
>> listed as an informative reference). Since unified-plan is not
>> expected to go to RFC, having a normative reference to it is problematic.
> Have to move this pointer to JSEP. I think JSEP has the necessary
> targets now.
Ok.

Thanks

-- Flemming


>> Thanks
>>
>> -- Flemming
>>
>>
>>
>> On 3/6/14, 8:25 AM, Flemming Andreasen wrote:
>>> I think the draft needs a bit more work in a few places (largely
>>> editorial and more text on some of the procedures). I'll sign up to
>>> do a pre-WGLC individual review to get you some more tangible and
>>> constructive feedback on that.
>>>
>>> Thanks
>>>
>>> -- Flemming
>>>
>>>
>>> On 3/6/14, 8:19 AM, Harald Alvestrand wrote:
>>>> I have uploaded an updated msid draft, without the msid-control:
>>>> section.
>>>>
>>>> As I said in the WG meeting, I think this is now uncontroversial, and
>>>> should be ready for WG Last Call.
>>>>
>>>> I request that the WG chairs consider whether they agree with this
>>>> opinion.
>>>>
>>> _______________________________________________
>>> mmusic mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/mmusic
>>>
>