Re: Blocker for PR: links to HTML5 spec

    Apology for letting this slip through earlier.

  Philippe's proposal of creating a reference section and pointing out its
dependency sounds good.
Some other options I can think of so far:
  1) Try to see if there is any stabler versions of the reference we can
find and move over there. (Though by looking at both
dev.w3.org and www.w3.org/TR, I don't expect we can find much.) And we leave
the rest of them as they are. It not the most
common practice but some RECs do refer to other less stable specs, the
Mobile Web Application Best
Practice<http://www.w3.org/TR/2010/REC-mwabp-20101214/> as
one of them.
   @Philippe, is it a hard requirement somewhere for a REC to point to other
RECs only? I am not able to find a pointer to that in
the Process Document <http://www.w3.org/2005/10/Process-20051014/>.

  2) If there is indeed concern of moving to REC with the less stable
reference from this WG, then maybe we can stay on PR until
we find ourselves comfortable moving forward. We can continue with bug fixes
and other stuff at the mean time.

 3) This might get me burn but I will bring it up anyway for the sake of
"creativity"... :-) The WHATWG version of the spec seems
to be more mature at this time and we can use some of those as references as
well. Still though, it's evolving as well.

   Among all these, option 2) seems to be a reasonable approach if we are OK
with staying in PR for a while.

cheers,
Zhiheng


On Thu, Feb 24, 2011 at 9:14 AM, Philippe Le Hegaret <plh@w3.org> wrote:

> On Thu, 2011-02-24 at 17:55 +0100, Anne van Kesteren wrote:
> > On Thu, 24 Feb 2011 17:44:28 +0100, Philippe Le Hegaret <plh@w3.org>
> wrote:
> > > On Thu, 2011-02-24 at 17:28 +0100, Anne van Kesteren wrote:
> > >> Also, I'm not really convinced that watering down all conformance
> > >> criteria is really the best approach here. If we care about
> > >> interoperability that
> > >> is.
> > >
> > > Why would we water down the conformance criteria? I don't think I
> > > suggested that. We already have tests and 2 implementations in fact and
> > > this discussion should have no impact on that at least.
> >
> > If you remove HTML as a dependency many important aspects will become
> > non-normative or undefined, in effect, right? Non-normative and undefined
> > material cannot be tested.
>
> Ah, but we would add the necessary bits in the specification before
> removing the normative dependency as part of an appendix. I don't know
> if that's feasible, thus my question to Zhiheng.
>
> Philippe
>
>
>
>
>