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

Re: [css-color] Editorial: have an exhaustive definition of the <color> type

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Wed, 25 Sep 2013 08:28:16 -0700
Message-ID: <CAAWBYDAfYOk_8nZaW4Kj0WYFesYKUJNWw9z677R9FhyAigNXNQ@mail.gmail.com>
To: Chris Lilley <chris@w3.org>
Cc: Simon Sapin <simon.sapin@exyr.org>, www-style list <www-style@w3.org>
On Wed, Sep 25, 2013 at 7:36 AM, Chris Lilley <chris@w3.org> wrote:
> Wednesday, September 25, 2013, 4:17:14 PM, you wrote:
>> The current ED has a section named "RGB Colors: the <color> type", and
>> following sections that define other color values like hsl(). This split
>> seems to imply that only RGB colors are <color>, the other values would
>> be something else.
>
> I agree that is misleading. Its not actually incorrect, since HSL is a
> reformulation of RGB. But I agree that the spec should distinguish
> between syntactic forms and semantic components. In particular, hsl
> values are a valid <color> type.

Yup, that's it - currently, the spec's model is that <color> is an
RGBA quad, and so it seemed okay to lump in the definition with the
"canonical" RGB-based syntaxes.

I'm fine with refactoring these sections to pull the definition out on
its own.  If we define multiple colors models, I'll *definitely* do
so.

> Also, as we will be adding other ways of specifying colours, it needs
> to be clear which are <color> and which are something else.
> Particularly as some of them include <color> as one part (for a
> fallback, for example).

They're all <color> - currently, they become an RGBA quad at
computed-value time.  That might change, of course.

~TJ
Received on Wednesday, 25 September 2013 15:29:03 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:34 UTC