W3C home > Mailing lists > Public > www-style@w3.org > April 2015

Re: W3C can say (?) what an ID means to CSS4, in the fast profile context

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Tue, 7 Apr 2015 16:11:11 -0700
Message-ID: <CAAWBYDDtt4ixSMeCuKhVkRt4x72bOaRfsG+m0yJyhsA3eZOC2Q@mail.gmail.com>
To: Peter Krauss <ppkrauss@gmail.com>
Cc: "www-style@w3.org" <www-style@w3.org>
On Tue, Apr 7, 2015 at 12:45 PM, Peter Krauss <ppkrauss@gmail.com> wrote:
> I will try to summarize...
> Facts when rendering XML or HTML documents with CSS:
> * Designers have the option to use CLASS and ID attributes, and the ID
> option is to express the uniqueness.
> * W3C need to listen to the design community, and statistics speak for them:
>    there are a lot of ugly documents with non-unique IDs... we must to
> tolerate non-unique IDs.
> in view of the facts ... There are two kinds of  "W3C workarounds", for
> tolerate non-unique IDs:
> 1) use the first. The DOM's tradition, with the getElementById(elementId)
> method, that
>     "must return the first element, in tree order", http://www.w3.org/TR/dom
> 2) ignore uniqueness of IDs. Is not an explicit W3C workaround,
>     but it is the (worst) practice in the CSS-parsing of all popular render
> engines (like Blink, Gecko or Webkit);
>     and, the W3C's "nothing to declare" is near to an agreement.
>     This practice say to designers "ID=CLASS" and
>     created a culture of "invalid IDs" in the content production.
> But now, with the CSS4's fast profile proposal, since 2013,
> http://www.w3.org/TR/selectors4/#fast-profile
> we can change this culture of "worst practices"...
> PROPOSAL (? is possible):  let's drop "repeated ID support" in the fast
> profile definition of CSS4.
>        http://dev.w3.org/csswg/selectors-4/#profiles
>    To do it explicitally, the CSS4 must express uniqueness as an
> extra-constraint for the fast profile.
> ... would be a gift, a W3C's bounty for designers and template-engines that
> produce non-ugly documents (docs with real IDs),
> and an incentive for render-engines really use distinct algorithms for each
> profile,
> looking for effective optimizations in the fast profile.
> PPKrauss
> PS1: there are many old discussions about ID in XML/HTML/CSS... Here an
> example as reference,
>    http://programmers.stackexchange.com/q/127178/84349
> PS2: about render engines, they work with "ignore uniqueness of IDs"
> strategy in the sense that
>   not use something like getElementById(), the CSS-render algorithms see ID
> attribute as see CLASS attribute.

There's no significant speed difference between matching all elements
with a given ID vs matching only the first, so I don't think there's
really a use-case for doing this.

As Boris said, the "fast profile" is just the set of selectors used
for CSS (rather than other Selectors-using APIs, like
querySelector()).  It's thus required to be a subset of the complete
profile; it can't have any actual behavior differences.

Received on Tuesday, 7 April 2015 23:11:58 UTC

This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:50 UTC