Re: On the way to CR for webrtc-pc

On Tue, Apr 18, 2017 at 2:05 PM Stefan Håkansson LK <
stefan.lk.hakansson@ericsson.com> wrote:

> On 17/04/17 12:01, Philip Jägenstedt wrote:
> > On Tue, Apr 11, 2017 at 2:33 PM Stefan Håkansson LK
> > <stefan.lk.hakansson@ericsson.com
> > <mailto:stefan.lk.hakansson@ericsson.com>> wrote:
> >
> >     Hi all,
> >
> >     as announced [1] our ambition is to request the transition to CR
> during
> >     this month. We received no negative feedback, so that is what we're
> >     working towards.
> >
> >     We've made good progress on the open Issues, and think we will be
> able
> >     to resolve them satisfactory before the end of the month. We're
> doodling
> >     for one more Virtual Interim just in case we need a meeting to
> resolve
> >     some issues (and it's early May so that would move us into next
> >     month) [2].
> >
> >     Since we in [1] asked everyone to file new Issues for everything they
> >     wanted addressed before asking for transition to CR, and that is
> about a
> >     month ago, we plan to label new Issues raised between now and the
> >     transition request "to be dealt with after transition to CR has been
> >     requested" or similar, and deal with them after the transition
> request
> >     has been issued.
> >
> >     Let us know if you think any of this is unreasonable.
> >
> >     Stefan for the chairs.
> >
> >
> >     [1]
> https://lists.w3.org/Archives/Public/public-webrtc/2017Mar/0063.html
> >     [2]
> >     https://lists.w3.org/Archives/Public/public-webrtc/2017Apr/0025.html
> ,
> >     http://doodle.com/poll/ecs5efy9r5f9747e
> >
> >
> > I'm not super familiar with what transitioning to CR entails, but can
> > you say something about what goals you have for interoperability, in
> > both the "implementations pass the same tests" and "two implementations
> > can communicate" sense) sense?
>
> I'm not super familiar either! But my understanding is that for moving
> to CR there is no real need to have done testing, or that
> implementations are available. To quote the process document
> (https://www.w3.org/2017/Process-20170301/#Reports):
>
> "A Candidate Recommendation is a document that satisfies the Working
> Group's technical requirements, and has already received wide review.
> W3C publishes a Candidate Recommendation to
>
> - signal to the wider community that it is time to do a final review
> - gather implementation experience
> - ...."
>
> However, as also said in the process document the CR may identify
> features as "at risk" if we think there will not be sufficient amount of
> implementation when it is time to move to PR. That's why we asked to
> that kind of input in
> https://lists.w3.org/Archives/Public/public-webrtc/2017Mar/0114.html


OK, so https://www.w3.org/2017/Process-20170301/#implementation-experience is
the bit I was looking for. To enter CR a WG "must document how adequate
implementation experience will be demonstrated", but it's the transition to
PR where the WG "must show adequate implementation experience."

> In my work to bring Blink's Web IDL files closer into alignment with the
> > specs that we link to, I'm mostly looking for non-standard things that
> > need attention, but I also notice things that are in specs but not in
> Blink.
> >
> > On RTCPeerConnection, there's at least currentLocalDescription,
> > pendingLocalDescription, currentRemoteDescription,
> > pendingRemoteDescription and canTrickleIceCandidates. There are also
> > things from partial interfaces that are harder to spot, like the sctp
> > attribute. Some of these are not in Gecko either.
>
> This is good information, but, as said, at this stage it is even more
> interesting to know for what features there is no *plan* to implement.
> With that information (from all browser vendors of course) we can mark
> those features "at risk" at this stage if we believe there will not be
> two implementations available when it is time to move to PR.
>
> >
> > Would it be helpful to the WG with a list of things that seem to
> > implemented in <2 browsers, or what criteria are you using?
>
> That would be very helpful, especially if each item comes with info
> saying e.g. "planned to be implemented" or "no plans for implementation".
>

I'll ask Mark Dittmer (cc'd) about how we could get this list. Since Edge
hasn't shipped WebRTC yet (right?) I think we could use the in-development
Web IDL diff tooling to find bits of the spec that don't exist in Blink and
Gecko respectively, and then compare the lists.

I couldn't say anything about implementation plans, though, not even for
Blink. If it wouldn't be useful without that, then I could hold off until
closer to PR, which I guess is when you'll want to start removing things
from the spec?