Re: [css-animations] When/how are keyframe values computed?

On Aug 8, 2014, at 9:23 PM, L. David Baron <> wrote:

> On Saturday 2014-08-09 01:20 +0000, Sylvain Galineau wrote:
>> On Aug 8, 2014, at 5:38 PM, Tab Atkins Jr. <> wrote:
>>> On Fri, Aug 8, 2014 at 5:35 PM, Sylvain Galineau <> wrote:
>>>> Yes. Also, keep non-animatable properties, switch non-interpolable values at half-way etc. (I'm still not clear on why we wouldn't do the latter for transitions but that's another topic).
>>> We do - the resolution to change the behavior of non-animatable
>>> properties was a general change to interpolation behavior, not an
>>> Animation-specific one, unless I'm remembering incorrectly.
>> That was my recollection as well, but I thought David remembered differently. Maybe I misunderstood him.
> specifically says (with emphasis as quoted):
>>   - RESOLVED: Make *animations* transition *all* properties. Unless otherwise
>>               specified, discrete properties take their starting values below
>>               50% timing function progress, and end values above 50% timing
>>               function progress.
> In the text leading up to that resolution, I had written on IRC to clarify a previous comment:
>>   <dbaron> (doesn't apply to transitions)
> My understanding at the time was that we explicitly decided this did
> not apply to transitions, and I still believe it should not.  I
> think the point of the transitions spec is to take changes from
> external sources and animate them when possible, and I don't think
> that makes sense for properties that can't actually animate.
> -David

Thanks for digging this up.

I find this confusing as I'm not sure how to explain to authors that they can animate all properties between keyframes but not in transitions. I thought the goal was to allow all properties to animate whether using css-animations or css-transitions i.e. everything is animatable but only some value types interpolate smoothly.

Unless there are good examples to show why it's harmful to do this in transitions I think we should reconsider the scoping of this resolution to css-animations.

Received on Sunday, 10 August 2014 23:09:43 UTC