RE: [css3-images] new side-or-corner keywords?

It's not a matter of technical complexity.  It's a model thing, and a logical complexity.  As it stands today, background is about the box (physical).  Additionally, the gradient is a type of image -- unaffected by writing modes, transforms, or any other of the new fun toys that aren't stable specs yet.

I think  it's an absolute horrid idea to intermingle these properties just because it's fun to mix the new chemicals in a beaker to see what happens.

I don't have the bandwidth right now to go into more detail, but the short version is:  adding this new interaction is a bad idea.  There have been a number of bad ideas in CSS that have had to be reserved, and people lament later.  I'm calling this one out before it happens -- do with that what you will.

-----Original Message-----
From: Tab Atkins Jr. [] 
Sent: Friday, August 12, 2011 3:32 PM
To: Brad Kemper
Cc: Brian Manthos; Florian Rivoal;
Subject: Re: [css3-images] new side-or-corner keywords?

On Fri, Aug 12, 2011 at 3:30 PM, Brad Kemper <> wrote:
> I still don't understand. It seems simple enough on the surface to be able to determine e.g. what the 'start' side is, translate it to a physical side and proceed as normal.
> Or do you mean implementation-wise it is hard to track the value throughout the background and into the image?
> So far this is only about affecting the direction within the gradient itself, and not about affecting background property values such as 'background-position'. Does that make it harder to track?

Indeed.  It doesn't seem particularly harder to do this than it would be to track logical keywords in any other property, like text-align.


Received on Friday, 12 August 2011 23:02:42 UTC