Re: [css-color-4] color() feedback

On Mon, Mar 10, 2014 at 8:57 AM, Lea Verou <> wrote:
> Syntax
> ========
> I was reviewing the color() [1] syntax recently. While the functionality is
> sorely needed, I'm not so sure the syntax is optimal. Here are a few
> examples based on the current syntax:
> color(var(red) tint(20%))
> color(var(green) blend(yellow 20%))
> color(lime blend(yellow 20%) shade(5%))
> I think the number of nested parentheses make it hard to read and
> understand, especially on longer expressions. Why not something like:
> color(var(red), 20% tint)
> color(var(red), tint 20%)
> color(var(green), blend with 20% yellow)
> color(lime, blend with 20% yellow, 5% shade)
> `with` in blend would be optional. The number of arguments would be the
> number of adjusters - 1. No particular order of tokens in each argument
> would be  enforced, as it seems with all the existing adjusters, it's not
> needed for disambiguation.
> I think something like this reads much more naturally, with no extra
> verbosity.

I'm open to something like this.  I mainly went with the current
design for the similarities with things like 'filter'.

> blend(a) adjusters
> ===================
> I'm pretty sure having both a `blend` and a `blenda` adjuster [2] is going
> to confuse the hell out of authors, many of which will expect `blend` to act
> like `blenda` and take all channels of the colors involved into account,
> including the alpha channel. I think it would make more sense if `blend` did
> what `blenda` is currently specified to do and `blenda` was dropped. Authors
> would still be able to use the current functionality of `blend`, with the
> `a` adjuster.

I'm basically copying SASS here.  Note that the blenda() adjuster does
*not* just blend the alpha channel along with the other colors; it
interprets the alpha as being an additional factor into the weighting,
so that blending a solid red with a mostly-transparent yellow won't
ever shift the color very far away from red - this is similar to how
compositing works.

It's possible that I should name these differently, because they're
very different operations.

> contrast adjuster
> =================
> As specified, the contrast adjuster [3] is based on a target contrast of
> 4.5. However, WCAG AA permits a contrast ratio as low as 3, for bold text
> above 14pt or text above 18pt [4]. If the contrast adjuster allowed for a
> <number> parameter to specify target contrast, it could cover this use case
> as well.

Are you talking about saying "I want X contrast, gimme that"?  Or "use
X contrast as the minimum, and otherwise act like normal"?

> I'm still not sure about whether this covers most use cases. I've
> extensively experimented with trying to find sufficient contrast ratios for
> color combinations (I've even built a tool about this [5]) and in many cases
> I knew whether I wanted a lighter or darker color and just adjusted the
> lightness until it passed WCAG. With the contrast adjuster as it stands,
> this is not possible to control.

Can you explain this in more detail?  I'm not quite sure I understand.


Received on Monday, 10 March 2014 21:36:49 UTC