Re: [css3-transitions] Interpolating transforms: a proposal to avoid Euler angles in favor of using quaternions

Sorry for the slow reply here.

The WebKit implementation of quaternions looks like it should produce 
equivalent rotations to the Gecko implementation.

That said, I think we should use Tim's original algorithm suggestion for 
the spec, primarily because it doesn't require any fudge factors. Not 
requiring any divisions either is always nice too.

- Matt

On 31/08/11 4:00 PM, Boris Zbarsky wrote:
> On 8/29/11 9:24 PM, Dean Jackson wrote:
>>> Now, I'm going to wimp out here and ask that Boris, the original 
>>> author, or someone else check this to make sure it is acceptable.
>
> Ccing Matt Woodrow, who did the Gecko implementation of this stuff. 
> I've asked him to look over the code you posted.  He probably has 
> other feedback on the spec too, having implemented it.
>
> We may also have a further proposal for handling interpolation of 3d 
> matrices with opposite determinants better, but it'll be at least a 
> month before that's likely to be written down in a coherent form.
>
> -Boris
>

Received on Sunday, 4 December 2011 22:22:22 UTC