Re: Publish a new Candidate Recommendation for Presentation API?

Hi Mark, all,

Le 22/11/2016 à 02:20, mark a. foltz a écrit :
> Francois, Anssi,
>
> I was wondering if we should set a milestone of publishing a new
> Candidate Recommendation, given the amount of work that has happened
> since the last one was published.  Please let me know what your thoughts
> are from a process point of view.

The Process requires the publication of a revised Candidate 
Recommendation as soon as substantive changes were made [1]. We resolved 
ambiguities in algorithms. Such changes fall into the "substantive 
changes" bucket and we did a number of them, so I believe the group 
needs to publish a revised Candidate Recommendation (CR).

Looking at milestones in the WG charter [2], we said publication as a 
final Recommendation would happen by end of June 2017 (Q2 2017). A spec 
must remain at the Proposed Recommendation (PR) stage during at least 4 
weeks, a bit more in practice. This means publication as PR would need 
to happen by end of May 2017. A spec must remain at the CR stage during 
at least 4 weeks as well. This means we need to publish a revised CR end 
of April 2017 at the very latest.

Things take longer in practice each time for various reasons (e.g. time 
needed to schedule a call with the Director, last-minute comment from a 
member during PR review, etc.).

With that in mind, I would say the group should aim for the publication 
of a final CR beginning of March 2017.

We do not have to wait until then though! We could typically publish a 
revised CR as soon as outstanding issues have been resolved. We would 
still have time to publish a third CR if we uncover new issues afterwards.

Two additional questions that I think the group should consider before 
publishing a revised CR:

1. The current CR does not single out any "at risk" feature (features 
that could be removed from the spec without publishing a new CR if they 
end up not being supported). Can implementers confirm that they plan to 
support all features defined in the spec?

2. Current CR exit criteria include "at least one implementation of the 
1-UA mode and one implementation of the 2-UA mode" for the receiving 
side. TPAC discussions suggested this criterion might need to be 
relaxed. Are we confident we'll have at least one 2-UA receiver 
implementation in particular?

Thanks,
Francois.


[1] https://www.w3.org/2015/Process-20150901/#revised-cr
[2] http://www.w3.org/2014/secondscreen/charter-2016.html#deliverables

Received on Tuesday, 22 November 2016 13:17:09 UTC