Re: [selectors-l4] a crazy observation?

Do you think custom pseudos with JS hooks will be less controversial to achieve and less risky perf-wise for implementors? I don't know much about the actual proposal so I can't really say. But from a long distance this doesn't obviously sound like an easier path to me, fwiw.

From: Brian Kardell <bkardell@gmail.com<mailto:bkardell@gmail.com>>
Date: Saturday, July 13, 2013 5:33 AM
To: "Tab Atkins, Jr." <jackalmage@gmail.com<mailto:jackalmage@gmail.com>>
Cc: www-style list <www-style@w3.org<mailto:www-style@w3.org>>
Subject: Re: [selectors-l4] a crazy observation?
Resent-From: <www-style@w3.org<mailto:www-style@w3.org>>
Resent-Date: Saturday, July 13, 2013 5:33 AM


On Jul 12, 2013 1:04 PM, "Tab Atkins Jr." <jackalmage@gmail.com<mailto:jackalmage@gmail.com>> wrote:
>
> On Fri, Jul 12, 2013 at 9:47 AM, Brian Kardell <bkardell@gmail.com<mailto:bkardell@gmail.com>> wrote:
> > That isn't what i am proposing at all though Tab.  I am merely proposing
> > that prioritizing a means for extension is time better spent than debating
> > whether we can or cannot implement fairly specific and potentially
> > problematic features in CSS...
>
> I don't believe the two impede each other in any significant way, and
> thus there's no need to punt one to get the other.  We don't do things
> linearly.
>
> ~TJ

There is nothing technically limiting, that wasn't my meaning.  In practice though we all prioritize our limited time so it isn't like there is no cost expense that doesn't affect the other.  I would be very happy to begin this either way though.  It seemed to me a few things added up:  1.  Qsa/css split is undesirable here 2.  It seems perf-wise no implementer has acceptable answers and don't want to risk 3. A few of these have been among the longest discussed and maybe controversial in some respects to about just what they are and how they should work.  I was simply suggesting that an enabling method to allow author extension may help on all of those points (and more) without setting a split precedent - and do so in a way which delivers author value faster and in a forward compat way.