Re: Methods of working: WebRTC WG and the PeerConnection spec

On Thu, May 12, 2011 at 20:46, Ian Hickson <ian@hixie.ch> wrote:
> On Thu, 12 May 2011, Harald Alvestrand wrote:
>>
>> now that we have the WEBRTC W3C WG, we need to figure out an expeditious
>> way forward for getting the specs we need done.
>
> Well, we have a spec already, so that part seems to be mostly done, modulo
> addressing implementation feedback.

Yes, that's one spec - we'll have to figure out if we need more or
not. But I trust that the WG will come to conclusions on that.

>> One possible course (but not the only one, of course):
>>
>> - Split out the PeerConnection definition and associated stuff into a
>> separate document (exactly what "associated stuff" needs to come with
>> it is something we have to discuss)
>> - Publish a snapshot as a W3C draft under the WEBRTC WG's banner
>> - Attempt to get focused review of that using the WEBRTC participants
>> and mailing list
>>
>> Does that make sense to you?
>
> I'm happy to prepare a non-normative extract much like is done with the
> WebVTT stuff:
>
>   http://www.whatwg.org/specs/web-apps/current-work/webvtt.html
>
> ...if that would help people review the relevant parts of the spec.
>
> If the W3C is interested in publishing such snapshots under the W3C patent
> policy, I'm sure everyone would agree that that would be a good thing.

I'm a bit worried about that - since, in my role as W3C WG chair, I'll
have to consider the standalone document normative, I see less chance
of troubles if it was considered a standalone document by all, the way
the Websockets API is handled (if I understand that correctly).

But I don't have strong opinions, I'll check with others to see what
their experience of the different methods has been.

> As far as where review comments go, I'm happy to take comments from
> wherever. For logistical reasons, I only promise to respond to all the
> feedback sent to the WHATWG list. The HTML working group at the W3C has
> been using W3C's Bugzilla instance to handle HTML spec feedback, which has
> also worked well; I'd be happy to handle PeerConnection feedback through
> that mechanism as well. I imagine Mike would be happy to set up a
> component for us if we want to do that. There's already a number of bugs
> on the topic filed there.

Bugzilla seems good to me for specific issues that have yes-or-no
answers; as you said in another discussion, bug-filing is good for
handling bugs, but not so good when we need to discuss approaches and
overarching issues - if such discussions occur on the WEBRTC mailing
list, I'm sure we can figure out how to summarize the conclusions (if
any) of that and pass it to the WHATWG list if that makes the
logistics easier.

I don't see a big problem with making sure comments get communicated.

Received on Friday, 13 May 2011 10:13:24 UTC