Re: pilot case?

We've been having some conversations on Twitter, I'd like to bring
them back to the list.  The (much smaller) group has discussed some of
this in the past, but I think we should iterate and have another go...
An idea can start as a draft - or even a wiki - or have a well
documented implementation.  How you go from there to 'getting close'
is kind of a much bigger topic, so I was suggesting something more
focused.

Essentially:  Can we (especially having members now from many
orgs/WGs) help define some de-facto sorts of practices that make a
smooth(er) process from concept to (possibly) standard.  A common
github layout that answers a few questions might be one thing that is
handy.. I was proposing it might be worth taking something that has
already been adopted by a WG and seeing how to come up with a standard
test runner, etc such that by the time implementors or the WG get
there, there is already something community provided (and possibly
partially reviewed by a few implementors) that they can pick up and
run with... Make the process easier.

In the 'larger picture' we might add advice like:  if it is aimed
toward w3c, 'use respec with unofficial draft and put it here' in the
layout... I think that would be a great suggestion.  Not sure what we
do if it is aimed at ECMA?

It's not that these need to be requirements, it just ideally should be
easy(er) to spin something up and review it too.  Having some kind of
test case that lets us practice, hone and get some experience with how
it might work seems like a decent idea... There have been pokes at
this by others - but nothing tremendously portable or advantageous to
the next person that comes along with an idea or wants to help move
the Web forward.


Thoughts?


On Mon, May 6, 2013 at 12:28 PM, François REMY
<francois.remy.dev@outlook.com> wrote:
> You’re right, we should probably make tests cases and hand them over to the
> CSSWG. They would appreciate to have us contribute to their test suite and
> show the prollyfills not only enable the authors to use the technologies
> more quickly, but also let the implementers implement them more easily.
>
> I guess your demo test page already contains a lot of test cases, now we
> should turn each of them into a separate test case (probably a reftest, this
> is the easiest option I think) and I can add them to the CSSWG mercurial (I
> already have an account on it).
>
> I may even work on this tonight, this doesn’t seem relatively
> time-consuming, the most difficult part is to write the assert statements in
> plain English ^_^
>
>
> De : Brian Kardell
> Envoyé : lundi 6 mai 2013 18:15
> À : public-nextweb@w3.org
>
> I worked this up over the weekend, it's just started but perhaps we
> can use it as a pilot case: It is comparatively simple and already in
> draft form with the CSS WG.
>
> https://github.com/bkardell/selectors-L4-link-prollyfills
>
> In other words, can we use this (or something else if the group
> prefers) to figure out how to make a process work?  Not just in terms
> of fork/evolve, etc - but in terms of making this a progressive
> process... Ideally as these get better, we would add test cases or
> examples that add value to the standards body (CSS WG in this case) so
> that, assuming a version _becomes_ a standard - those things are
> already contributed, agreed upon by a number of implementers, etc.
>
> --
> Brian Kardell :: @briankardell :: hitchjs.com
>



-- 
Brian Kardell :: @briankardell :: hitchjs.com

Received on Tuesday, 7 May 2013 17:05:49 UTC