Re: [CSSWG] Minutes Telecon 2014-03-12

On Thu, Mar 13, 2014 at 12:00 PM, Tantek Çelik <tantek@cs.stanford.edu> wrote:
> In the spirit of adding something more productive to the conversation,
> all mentions/pressures of shipping aside, it does seem prudent for the
> CSSWG to make efficient progress on this.
>
> Tab,
>
> Is there an ordered-by-proposer-preference 1..n list of proposal
> summaries (combinator(s) and/or pseudos) for Shadow DOM selection?
> (URL?)
> (including of course an "n. insert your own new proposal here" just as
> a catchall of course)
>
> I trust your experience with working on this to provide such a list,
> and with such a list reflecting a reasonable amount of consideration /
> expert opinion for what are better/worse solutions (and why, even if
> it's just your personal preference) - I see that as more informative
> than just "can live with".
>
> I honestly had trouble following some of the variants being tossed
> around on the telcon (especially some of the new proposals that seemed
> to get brought up on the call itself).
>
> I feel like if the WG were to see such an ordered list, then it may be
> easier (faster?) to build/gain consensus on a choice among them.

The current state of the spec reflects me and fantasai agreeing on things:

* A ::shadow pseudo-element, which represents the ShadowRoot object
(the contents of the shadow tree are treated as children).  Replaces
/shadow/.
* A ::content pseudo-element, which represents the <content>
distribution list itself (the elements distributed to the <content>
are treated as children).  Replaces /content/.
* Renames /shadow-deep/ to just /deep/, as the connection with
/shadow/ no longer exists, and it's not *really* about shadows, but
rather about being a super-descendant combinator.
* Keeps :host/:host().
* Renames :ancestor() to :host-context(), as it's a better name and
has a much better visual connection with :host(), which it's very
similar to.

This appears to be acceptable to both fantasai and me (and by
extension, the Blink team).  By agreeing with fantasai's position in
several places, it also seems to address the concerns of many other
people in the WG, who agreed with fantasai in preferring
pseudo-elements over combinators.  Hopefully this is a more harmonious
solution in general; we should discuss this at the next telcon.

~TJ

Received on Monday, 24 March 2014 23:52:27 UTC