Re: Animation restart

Frederik Elwert:

>So my questions are:
>
>      * Is my original approach valid at all? 

Yes, however due to accuracy problems and because it is
simpler for current viewers, it is often more practical to
use for both animations begin="start.click" end="stop.click".
Anyway, there should be no problem with syncbase-values,
if a browser/viewer has implemented, what is defined in SMIL ;o)

(Another option to simplify things is to use no style attribute and to
note everything in presentation attributes. With this already tiny-viewers
should manage the sample including animation, not just full-viewers -
but this simplification is not much related to your animation issue.)

Typically viewers have less bugs for begin than for end and at least
Opera 9.50-11.00 has not so much bugs in timing issues at all - 
from my timing tests 10.60/11.00  pass already 110 from 168 timing 
tests for the tiny profile. 
But your example looks indeed interesting to explore the behaviour
in slightly more complex situations than basic tests.


>      * Which behaviour is the defined one? If any? 
That what is described in SMIL. Of course you can use syncbase-values
and because you note no special restart conditions or negative offsets
(some viewers have problems with) or other more advanced complications, 
it should work as expected with proper implementations.
About fill="freeze" - the value at the end of active duration is frozen,
what is not necessarily the last value in the values list.
Due to my tests typically Opera and the Adobe plugin have a good
understanding of the fill attribute, Batik/Squiggle fails in all of my tests
and WebKit passes some tests.

>      * Are there better solutions?
Help implementors to fix bugs ;o)
Until there are proper implementations available, try to work around
bugs, you can simplify as suggested above.

Received on Thursday, 6 January 2011 13:40:00 UTC