W3C home > Mailing lists > Public > www-style@w3.org > March 2011

Re: [css3-images] conical-gradient()?

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Tue, 22 Mar 2011 11:08:07 -0700
Message-ID: <AANLkTi=P2-JTqqO13zZUBAxt183qMyc4x8nw3-6B1mNi@mail.gmail.com>
To: Rik Cabanier <cabanier@gmail.com>
Cc: Stephen Hay <haymail@gmail.com>, Lea Verou <leaverou@gmail.com>, www-style@w3.org
On Tue, Mar 22, 2011 at 10:57 AM, Rik Cabanier <cabanier@gmail.com> wrote:
>> > I think it would be better to call this an angle gradient instead of
>> > conical.
>> In Photoshop it is indeed often referred to as angle gradient (and
>> this is technically correct), although most designers I know
>> understand the term "conical". I actually think conical best describes
>> what the gradient does visually.
> A conic radial gradient also resembles a cone. If we call them angle
> gradients, there is less chance of confusion.

I don't much care about the name.  We can bikeshed over that in a year
or two when I start work on Image Values 4.

>> > Conical gradients are just a special case of a radial gradient where the
>> > midpoint of start and ending circle is different.
>> > This seems like an edge case. Neither SVG or Illustrator implement such
>> > a
>> > construct so it must not be that common.
>> I would call that a hasty generalization. Photoshop *does* implement
>> it, so making a similar generalization one could argue that since
>> there are more Photoshop users, it is very common.
> Are you advocating that we also add diamond gradients?

FWIW, I wanted to add square and diamond gradients.  You would just
pass in different keywords instead of 'circle' or 'ellipse'.  However,
creating them would require some hacking in graphics, like creating
four linear gradients and masking them together.  I decided they were
niche enough to not worry about them for now.

(If I added them today, I'd do them as a separate function, since you
can't interpolate from an elliptical to diamond gradient in such a way
that the intermediate forms can be sanely represented in CSS, or
efficiently constructed using platform graphics libraries.)

> At one point I suggested adding focal point support to radial gradients.
> This is a very common idiom which is supported by the WebKit gradients and
> core graphics but it was deemed to much of an edge case.

Indeed, and I still believe that to be true.  I'll revisit the issue
in Image Values 4.

Received on Tuesday, 22 March 2011 18:08:59 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:44 UTC