W3C home > Mailing lists > Public > www-style@w3.org > July 2012

Re: Property interactions and computed style

From: L. David Baron <dbaron@dbaron.org>
Date: Thu, 12 Jul 2012 17:55:47 -0700
To: "Tab Atkins Jr." <jackalmage@gmail.com>
Cc: Simon Fraser <smfr@me.com>, "www-style@w3.org Style" <www-style@w3.org>
Message-ID: <20120713005547.GA10408@crum.dbaron.org>
On Thursday 2012-07-12 13:53 -0700, Tab Atkins Jr. wrote:
> On Thu, Jul 12, 2012 at 1:44 PM, Simon Fraser <smfr@me.com> wrote:
> > There are a couple of places in the CSS3 transforms spec where property interactions cause the normal behavior of one value of a property to be affected by another property.
> >
> > For example, we now say that "background-attachment: fixed" behaves like "background-position: scroll" if the element is affected by a transform. Similarly, "transform-style: preserve-3d" behaves like "transform-style: flat" if opacity is non-1.
> >
> > In such cases, what should the computed value of the affected property be?
> 
> I don't feel strongly about the final answer, but it makes sense to me
> to have the "behaves like" be reflected in the computed value.  When
> these sort of interactions are purposeful, that's usually what we do.

I'd also like to ensure that the computed value of a property can be
computed based on local data:  that is, based on the specified style
on the element and the computed style of its parent.  I think
deviating from that makes CSS much harder to implement with
algorithms that execute in parallel, which I think is quickly
becoming more important, since CPUs aren't getting faster, we're
just getting more cores.

(I think it would be great for getComputedStyle to be affected by
these transformations, though, but I think it's important that
inheritance not be, and the "Computed Value" lines we have describe
inheritance.  We probably actually need to start adding "Used Value"
lines that describe what getComputedStyle returns.)

Therefore, for the background-attachment case, I think, strongly,
that the computed value should not be affected.

In the transform-style case, I think you're talking about 'opacity'
on the element itself, so I don't feel so strongly, but I tend to
prefer not affecting the Computed Value.

-David

-- 
𝄞   L. David Baron                         http://dbaron.org/   𝄂
𝄢   Mozilla                           http://www.mozilla.org/   𝄂
Received on Friday, 13 July 2012 00:56:49 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:56 GMT