Re: <main>

What I meant was that I would (at first blush and with no knowledge of prollyfills) extend the current shims that provide support for header/footer/article etc. Extend it so that <main> is recognized and add the hooks for ARIA and keystroke capture etc.

Having said all that - I completely agree with Brian's points below. And with that I see this _could_ be a prollyfill solution. I will say that I stand corrected in that this for sure is prollyfill by definition. My point about polyfill is simply as I said - given no prollyfill tooling I'd extend the polyfill that already exists.

Does that make sense while providing no benefit? :)



On Nov 28, 2012, at 9:06 AM, Brian Kardell <bkardell@gmail.com> wrote:

> 
> 
> 
> On Wed, Nov 28, 2012 at 10:49 AM, Clint Hill <clint.hill@gmail.com> wrote:
> Sorry - my wording was bad. 
> 
> I agree it's a prollyfill. I think however the "solution" is already a polyfill. Which is why I guess I need more opinion. 
> 
> 
> 
> Sorry, I don't understand.  For posterity and in case anyone else is not on the same wavelength - can you explain?  Particularly, I am suggesting that prollyfills should always exist outside the native namespace, thus <x-main> which is just simple x-tags stuff.  I believe/advocate/propose that all new tag proposals should follow a Web Components/prollyfill model.  Take a look at the arguments that have happened around <article> and <header> etc and how they came about... Everyone has the best of intents and I believe is acting in good faith - but clearly people misunderstand.  The real potential argument that I can see against this is that one man's <x-main> might not mean the same as the others as it would be based on the prollyfill - but I think that in hitch we have some decent answers to that.  For anyone not familliar, we made Hitch before implementers decided to go with <x- and while the solution agreement still seemed to be attribute based, so we do data-widget and that property contains a uri capable of identifying not just the particular web component, but the versioned implementation as well... Given this kind of approach you could have something like:
> 
> <x-main data-widget="http://www.foobar.com/mywidget.js#1">
> 
> which gives you something like a namespace.   Given that much information, there is really nothing that couldn't slowly gain traction even to the point of being respected by screen readers/search engines early if it gained a wide enough audience before being formalized.  In other words, if version #6 finally had iterated enough and had both a mature draft proposal and a lot of adoption, you could effectively submit it and people could essentially just alias the two in their implementations... 
> 
> I'd like to mention that all of this is mostly a way to get some of these conversations that we need to have rolling:  What kinds of things should we prollyfill - what kinds of principles go behind them, how does prollyfill progress through to maturity and eventual submission, etc.
> 
>  
> Brian Kardell :: @briankardell :: hitchjs.com
>