W3C home > Mailing lists > Public > www-style@w3.org > April 2010

Re: [css3-transitions] Adding keyframes to transitions

From: Brad Kemper <brad.kemper@gmail.com>
Date: Wed, 7 Apr 2010 11:39:34 -0700
Message-Id: <38AD0886-0DBB-47B9-9C16-E42E619C3601@gmail.com>
To: "Tab Atkins Jr." <jackalmage@gmail.com>
Cc: Perry Smith <pedzsan@gmail.com>, www-style list <www-style@w3.org>
On Apr 7, 2010, at 10:20 AM, "Tab Atkins Jr." <jackalmage@gmail.com>  
wrote:

> On Wed, Apr 7, 2010 at 10:13 AM, Brad Kemper <brad.kemper@gmail.com>  
> wrote:
>> Aside: I think if the bounce involved 'left', then it would be  
>> overridden by
>> the 'left' in the transition-property. At least, that makes the  
>> most sense
>> to me.
>
> Really?  I would think the opposite.  The reason I'd like the
> animation to override the transition is so I can express greater
> control over the transition.
>
> For example, say all I want to do is transition 'left' on an element,
> but I want it to do so by making it quickly fly from start to end, and
> then bounce slightly before settling down.  You can't express this
> with any timing function (nor will you ever be able to, since it's
> non-invertible).  Potentially, I could do this by creating an
> animation that does this, and then hooking it to a 'left' transition.

I'm OK with that line of reasoning as long as it is doable. I mainly  
fealt we should say one way or the other, and wasn't sure if we had yet.

My reason for going the other way was that the transition is based on  
the changed value of 'left'. If you say that the inner animation  
overrides that, then there is no longer a left value to transition. I  
think what you realy want is to multiply the two together for the  
final animated effect, and that's more like what you describe. That  
would seem to complicate computed values, but works for me.

> To do this *properly*, of course, requires some way to refer to the
> start and end values in an animation, and do math to them.  You can't
> do that right now, but I can see it being added in the future, and
> wouldn't want to shut down the possibility of doing what I described
> above.

I'm not sure I follow, but that's OK.

> For now, since it's not very useful to manipulate the property you're
> transitioning, I'd just recommend not doing so.  I don't think it's a
> problem right now to say "you probably don't want to do that", and
> then later make it useful to do it.

We should define what it does. 
Received on Wednesday, 7 April 2010 18:40:24 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:26 GMT