Re: WebIDL roadmap?

Hi Dom, Robin.

Dominique Hazael-Massieux:
> > What are the plans for stabilizing WebIDL and pushing it to Rec? The
> > charter has it initially at CR in Q4 2008 - which obviously hasn't
> > happened :)

Apologies for not having had time to work on it recently.  My (most
recent) plans were to have it ready for LC some time around now,
basically before HTML5’s LC.  The number of open issues isn’t too great,
and I can certainly make an effort to resolve these within a few weeks.
(The last spurt of work I got done was while at the last SVG F2F, and we
have another one of those coming up in a week. :))


Robin Berjon:
> In the case in which WebIDL is not realistically believed to be
> moveable forward quickly, I'd also like to ask the editor and other
> stake-holders what they would think of publishing a v1.0 consisting
> only of what is considered solid (assuming that is enough for
> referring documents to work with) and push some of the more complex/
> contentious/unfinished features into a v1.1 (work on which can be
> continued in the same way)? I know it would add a bit of overhead,
> but it would be really useful if we could have an anchor in the
> dependency game.

Right, so there are parts of the spec that are pretty solid (e.g. how ES
values get converted into IDL booleans is pretty uncontroversial) and
those that are more speculative (how multiple inheritance of interfaces
is mapped to an ES prototype chain, sequences/arrays, exactly how name/
index getters & setters are to be implemented).  Some of the fancier
features are exactly what some spec authors are referencing Web IDL for,

So what maturity do people need Web IDL to be in relatively soon?  If
it’s just LC, then that’s do-able.  I having a feeling though that there
could be substantial changes in response to LC comments, so that might
delay CR somewhat.  I think some people are holding off on doing a
thorough review of the spec until it is in LC, so I think it makes sense
to hurry up to get to that state soon.

I’m open to suggestions on how to proceed.  Splitting the spec into
parallel 1.0 and 1.1 versions could be done, but not without a little

Cameron McCormack ≝

Received on Wednesday, 16 September 2009 23:57:38 UTC