W3C home > Mailing lists > Public > www-style@w3.org > June 2013

Re: [css3-cascade] Editorial comments

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Mon, 10 Jun 2013 16:19:27 -0700
Message-ID: <CAAWBYDCxVcJ6NnNxe8jxw4s6NiCkfumzpgEwh8stwJfsw6vwRw@mail.gmail.com>
To: Simon Sapin <simon.sapin@exyr.org>
Cc: www-style list <www-style@w3.org>
On Sat, Jun 8, 2013 at 7:57 PM, Simon Sapin <simon.sapin@exyr.org> wrote:
> When listing "The precedence of the various origins", §4.2 should link to
> §4.2.1 and §4.2.2 to define origins and '!important', respectively.


> §4.2.1. Cascading Origins should define that imported stylesheets have the
> same origin as the stylesheet that imported them.


> Thorough the document, the word "origin" (especially when preceded by
> "same") should be disambiguated by linking either to §4.2.1 when it refers
> to cascading, and some other definition (maybe the WHATWG Fetch spec?) for
> the same-origin policy.
> Similarly, "imported stylesheet" should link to the section defining
> '@import'.


> With inheritance, the *computed* value of the parent becomes the *specified*
> value. This requires one of two things to keep the whole thing well-defined,
> although this is probably not a problem in practice:
> 1. The process of resolving a specified value into a computed value (which
> is specific to every property) must be idempotent.
> 2. Or §5.2 should say that the specified value becomes the computed value
> as-is if it was inherited from a computed value.
> I’d prefer 2, as 1 is a burden for every other CSS spec.

We resolved some time ago on the current wording:

> In §5:
>     Finally, the computed value is transformed to the actual value
>     based on constraints of local environment.
> s/computed/used/


> The first two paragraphs of §5.2. "Finding the computed value" should be a
> non-normative design principle, since the process for finding the computed
> value is normatively defined by every property. Considerations of absolute
> vs. relative value can not always be generalized (an unitless number is
> absolute in 'opacity' but relative in 'line-height') and some relative
> values are not resolved until layout as used values.
> Similarly in §5.3: the used value can differ from computed value more than
> just resolving remaining relative values. This process is specific to each
> property, and this spec should not generalize too much.

I think it's clear that this is talking in general principles.  The
actual definition clearly defines that the process is defined in each
property's definition.

Received on Monday, 10 June 2013 23:20:17 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:31 UTC