Re: [css3-transitions] Interpolating images

On Tue, Nov 30, 2010 at 1:35 PM, Robert O'Callahan <robert@ocallahan.org> wrote:
> On Wed, Dec 1, 2010 at 10:30 AM, Simon Fraser <smfr@me.com> wrote:
>> Let's not start assuming that GPU acceleration is available everywhere.
>> It's
>> going to be a few years before GPU acceleration is widely available on
>> mobile devices,
>> and different browsers have different levels of support for GPU
>> acceleration.
>
> I'm just saying, let's not design new CSS features around implementation
> characteristics that are likely to go away in the medium term. Especially
> performance characteristics.
>
> (Plus, animated image size changes are already common, and engines already
> need/want to be able to do them fast.)

While we're on the subject, how about relaxing the restriction on
image-orientation that it only be multiplies of 90deg?  By the same
argument, implementations already have to handle this with rotate
transforms, and they'll want these to be fast.  The only difference is
that image-orientation is layout-affecting, so I'd just define the
size of the image to be the minimum-sized axis-aligned bounding-box
around the rotated image.

~TJ

Received on Tuesday, 30 November 2010 21:59:02 UTC