Re: Talk on radial gradients

On Tue, Sep 1, 2009 at 2:42 PM, David Perrell<davidp@hpaa.com> wrote:
> Tab Atkins Jr.:
> | http://www.xanthir.com/test-gradient.php?shape=skewellipse&width=2
> | 00&height=200&start=ffff00&end=0000ff
> | http://www.xanthir.com/test-gradient.php?shape=quarter-circle&widt
> h=200&height=200&start=ffff00&end=0000ff
>
> #1: yellow spotlight shining into a blue fog
> #2: solar corona (expected of a simple radial gradient)

I'm gonna take that as a vote for #2.

> | I'm trying to decide if being able to specify the starting-position is
> | useful or not.
>
> Start offset from ellipse position surely useful for someone.

I'm sure it's useful for *someone*.  The question is if it's useful
enough to complicate the syntax and explanation with.  Basically,
starting-position and <length>-in-color-stop are incompatible, because
the former skews things and distorts absolute distances.  I either
drop starting-position, or limit you to using % in radial color-stops.
 I'm opting for the former, as it'll keep radial-gradient() closer to
linear-gradient and make it *much* easier to explain without dropping
into formulas.

> Plus something like 'reverse', like looking at the backside of the gradient?

I'm not sure what you mean; can whatever this is be accomplished by
reversing the color-stops?

> And let's not neglect an optional eccentricity variable for each color stop. E.g.:
> http://hpaa.com/csstest/radgrads.htm

I really hope you're joking.  ^_^

> (BTW, I found no compelling argument for tilted elliptical gradients.)

Cool.

~TJ

Received on Tuesday, 1 September 2009 21:26:03 UTC