Robert O'Callahan wrote:
On Fri, Aug 14, 2009 at 11:21 AM, Christopher Robert Jaquez <crjaquez@gmail.com> wrote:
Robert O'Callahan wrote:
On Fri, Aug 14, 2009 at 10:21 AM, Christopher Robert Jaquez <crjaquez@gmail.com> wrote:
I like that.  And for a linear gradient, the end value could
optionally be <angle> <color> <length>
 
That would just be <angle> <color>.

Not necessarily.  I intended <angle> <color> <length> as that is what is required to determine the point.  Otherwise, you have the direction but where do you stop?  The stop point is essentially a polar coordinate centered at the start point in this case.

Yes, but specifying a length here isn't very useful for authors.

A better approach would be to specify the <angle> *and* a point through which the end line must pass, which is what we're talking about on the other thread.

Why would a length not be very useful for authors?  In any case other than a gradient that goes all the way across the background, I am much more likely to know how wide I want the gradient to be than the coordinates of a point where I want it to stop.