Re: [whatwg] (X)HTML + SMIL?

On Sat, 27 Dec 2008, Bjoern Hoehrmann wrote:

> I note by the way that the "CSS Transitions" still lacks a complete de-
> finition of exactly when a transition begins, and from which old state
> to which new state it should go, and what happens once it has reached
> it. For example, from the incomplete and informal suggestions it seems
> that changes trigger transitions, and transitions trigger changes; that
> process does not terminate by itself. 

Yes. We're aware that this is terribly under-specified.

And it is quite complicated, since there are so many things to consider,
and CSS doesn't really say itself when *exactly* style updates happen.

Any advice you have here is appreciated.

> Similarily other changes may occur
> while a transition is in progress, what happens then is not clear at all
> absent such definitions.

We're clarifying what happens when the values of a transition change mid-
transition. If they are changes to the properties of the transition (say,
extending the duration) then they have no effect: the values are snapshotted
when the transition starts. If they are changes to the property that is
being transitions, then a new transition starts to the changed value from
the current value (which will be part way through a transition). Some people
want the ability to auto-reverse a transition, but we're not quite ready
to propose that at the moment.

Dean