Re: frozen value for discrete animation

Hello,

an additional questions about this and my case 1).


What I understand now is that fill="freeze" is defined in SMIL2 in
another way as in the 'SMIL Animation Recommendation 04-September-2001' 
or in SVG (1.1 or CR Tiny 1.2, still all using the old definition)


Old definition:
'freeze
The animation effect F(t) is defined to freeze the effect value at the last 
value of the active duration. The animation effect is "frozen" for the
remainder of the document duration (or until the animation is restarted - 
see Restarting animations).'

Together with the time interval model it is pretty clear, that the last value
for example in case 1) is 1 and not 2 because of the exclusive end of the
interval model, therefore my interpretation is consistent with the old
SMIL Animation Recommendation and with all SVGs, using this definition.

SMIL2 notes (timing module):
'freeze
Specifies that the element will extend past the end of the last instance of
the simple duration by "freezing" the element state at that point. The 
parent time container of the element determines how long the element is 
frozen (as described immediately below).'

and before:
'For algorithmic media like animation, the value defined for the end of the
last instance of the simple duration should be used.'
 
Well if we ignore the wording problem with 'simple duration' and 'active
duration' as already mentioned in the specification itself, the conclusion
will maybe the same, but this was simpler to decide for the old SMIL 
Animation Recommendation.
Because the time interval model is exclusive the time of end, still in case 1)
1 is the last instance of the active duration and not 2 (3 would be the last
instance for the simple duration, possible too with the current wording).

But in the animation module we find (3.3.5) an additional rule:
'If AD is not an even multiple of the simple duration d, f_f(t) = f_i(t), 
where i = floor(t/d).'

This would be applicable for case 1) too and we get a different result as
before,because AD < d, f_i(AD) is defined and has the value 2 and not 1
anymore, because AD is already the initial inclusive time of the second 
sub interval.

If we assume, that this specific rule in the animation module overwrites the
general rules in the timinig module (which seem to be never applied then for
animations), we get a deviating behaviour in SMIL2 (deviating from the
behaviour defined in the old SMIL Animation Recommendation and in the 
SVGs) for frozen discrete animations. 

Is this observation correct?



Olaf Hoffmann

Received on Thursday, 22 March 2007 15:22:28 UTC