Re: Length of LC comment period [Was: Ready for LC on the various drafts I edit]

On Sun, 6 Dec 2009, Arthur Barstow wrote:
> On Dec 4, 2009, at 7:19 AM, ext Ian Hickson wrote:
> > On Fri, 4 Dec 2009, Arthur Barstow wrote:
> > > 
> > > If we already have multiple implementations of a spec, I think the 
> > > spirit of the Recommendation track process suggests a shorter LC 
> > > period (say 2 months given the time of the year) and then (assuming 
> > > no substantive comments) moving the spec to Candidate.
> > 
> > I don't realistically think I'd have time to address the likely volume 
> > of comments in two months, since I am also dealing with HTML5's last 
> > call comments in the WHATWG and will likely be starting work on some 
> > more specifications in January.
> 
> Expediting the LC and CR could be important for dependencies. The only 
> dependency I know of without searching is the Widget Interface spec 
> (which is on schedule to enter CR in Jan) and it has a normative 
> dependency on Web Storage.
> 
> Thus, another option is to stagger the LC comment periods. For example, 
> 1-2 of the more "baked" specs could have a shorter LC period e.g. 2-3 
> months, the medium baked specs 3-4 months, etc. Is something like that 
> workable for you, in particular making the exit of LC for Web Storage in 
> 2-3 months?

Given the storage mutex problem, I would estimate that Web Storage is 
actually the draft that needs the longest review period. Going to CR in 
July would already be a huge acceleration on the schedule I have 
previously predicted, which was that these drafts would reach the CR stage 
in 2012. I'm very uncomfortable with accelerating the process even 
further; I think it would do a diservice to the Web.

Process-wise, Web Storage being in LC is not a blocker to the Widgets 
specifications going to CR or even REC, so I do not believe that 
expediting the process is necessary for dependencies. In fact, I would 
argue that the opposite is true -- if we go to CR in two months, then have 
to go back to WD, LC, and CR again three times over the next two years, 
the effect on the Widgets specifications will be much more disruptive than 
if we spend 6 months or a year in LC before going to CR, and then stay 
there until the test suite and implementations are done.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Sunday, 6 December 2009 13:35:15 UTC