Re: Deliverables Re: TPAC

On Jul 26, 2013 2:50 PM, "Yehuda Katz" <wycats@gmail.com> wrote:
>
> But the polyfills themselves have supported browsers!
>
> And prollyfills are completely new features. Do they want to include
every old prollyfill?
>
> On Jul 26, 2013 11:26 AM, "Marcos Caceres" <w3c@marcosc.com> wrote:
>>
>>
>>
>>
>> On Friday, July 26, 2013 at 4:34 PM, Yehuda Katz wrote:
>>
>> > caniuse is about whether things already work in browsers. This would
be about documenting p(r)olyfills (and what browsers they worked in).
>>
>> Each API they list also contains a "Resources" section that more often
than not has a link to a Polyfill.
>>
>> Have a look at any API there:
>> http://caniuse.com/#feat=matchmedia
>> http://caniuse.com/#feat=classlist
>>
>> And so on….
>> > It would create another dimension to caniuse, but if they were up for
it, sure.
>>
>> I think we can just send pull requests for the data tables they are
using:
>> https://github.com/fyrd/caniuse
>>
>> For example, classlist.json has:
>>
>> links:[ {
>> "url":"https://github.com/eligrey/classList.js",
>> "title":"Polyfill script"
>> }]
>>
>>
>>
>> --
>> Marcos Caceres
>>
>>
>>

I think we are conflating two things here... Things would have to graduate
to caniuse as it exists now at least it is for things with w3c drafts and I
think generally at least an experimental impl by some vendor.  We are
talking about fills for things of all levels of proposal maturity here
where they can compete a bit.

For example, hitch has fills for things in the selectors level 5+ wiki,
there isn't even a draft yet - and things for which there isn't even that
level of w3c commit yet.

There should be a place to explore that sort of thing, measure maturity,
cross-browser ability and other metadata as we have discussed.  If caniuse
wants to offer to collab, host, etc - that could be cool, but I think we
have to be clear what we are taking about here