Re: [rtcweb] Proposal for a JS API for NoPlan (adding multiple sources without encoding them in SDP)

[Peter Thatcher <[e-mail]>]

If Node.js wants a cleaner API, it's can certainly create one.  It's not at
all constrained by what goes in the browser or what the WGs decide.  In
fact, that would be an interesting way to explore a cleaner API to see how
well it works.


On Wed, Jun 26, 2013 at 5:19 PM, Bernard Aboba <[e-mail]>wrote;wrote:

> I do think No Plan represents progress toward containing the SDP monster
> in WebRTC 1.0, particularly for video scenarios. However, I also agree with
> Robin's arguments against SDP and O/ A, particularly for a server-side API
> (e.g. Node.js modules) where the current API makes no sense at all. So my
> perspective is to contain the monster in 1.0, then go with 2.0 first on
> server, then on the browser.
>
> On Jun 17, 2013, at 10:34 PM, "Peter Thatcher" <[e-mail]>
> wrote:
>
> I like your comparison to the data channels.  As I just pointed out in
> another email, I think it was good that we "contained the SDP monster", as
> you put it, with createDataChannel.  One of the purposes of
> createLocalStream/createRemoteStream is to allow a JS app, if it chooses,
> to "contain the SDP monster" when adding media streams.  It would still be
> needed for setting up the PeerConnection's transport (a monster container
> for a future day perhaps), but that's still significant progress in my
> book, and it does so with simple additions to the PeerConnection that don't
> attempt to blow up the WG.
>
>
> On Mon, Jun 17, 2013 at 5:14 PM, Bernard Aboba <[e-mail]>wrote;wrote:
>
>>
>>
>>
>> 2) Dissadvantages of using SDP in WebRTC.
>>
>>
>> Roman said:
>> "An unmanageable monster was created which currently stays in the way of
>> developing new functionality (bundle), building applications (does not
>> provide obvious ways to implement obvious tasks, like adding an extra
>> stream without re-negotiating all the existing ones) and even interop with
>> existing SIP endpoints (which was the original but now is complicated since
>> it would require a non trivial set of constraints and subsequent SDP
>> manipulation)."
>>
>> [BA] Hard to argue with this, but I would point out that by far the
>> ugliest part of the monster is the video hindquarters.  While one could
>> argue that we have been living with the warts of SDP for audio and
>> therefore know the workarounds, with video there are substantial
>> interoperability issues, *even among vendors who utilize the same
>> codec*, sometimes even in relatively basic scenarios (e.g. P2P video call
>> with H.264/SVC).  So the "multivendor legacy of interop" just doesn't exist
>> yet (at least, based on standards).
>>
>> So as I see it, it does represent progress that we have contained the SDP
>> monster's impact on the  the data channel, and I welcome Peter's effort to
>> enable applications who don't care about SDP to minimize its usage even if
>> it is not eliminated entirely.
>>
>> _______________________________________________
>> rtcweb mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/rtcweb
>>
>>
>