Re: [css3-2d-transforms] rotation and animation

2009/3/23 mozer <xmlizer@gmail.com>:
> 2009/3/23 Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>:
>> Hello,
>>
>> this comment applies to css3-3d-transforms as well.
>> http://www.w3.org/TR/2009/WD-css3-2d-transforms-20090320/#animation
>> http://www.w3.org/TR/2009/WD-css3-3d-transforms-20090320/#animation
>>
>>
>> In the related section about the transformation functions it is noted, that
>> for the <angle> type CSS Values and Units apply:
>> http://www.w3.org/TR/css3-values/#angles
>> This notes:
>> "Angle values should be normalized to the range 0-360deg
>> by the user agent."
>>
>> 2) authors might want to have a rotation from 30deg to 750deg,
>> what means two complete turns and not no animation,
>> or maybe from 90deg to -90deg, what is different from an
>> animation from  90deg to 270deg (what has a
>> different rotation direction too).
>> Concerning animation I think, it is pretty useful not to
>> modify the given values because this results in quite
>> different effects as typically intended by the author,
>> respectively authors cannot get the intended animation
>> effect with only one rotation function at all.
>>
>
> So you might want to use the 'turn' unit for this purpose as specified
> in the same place ( http://www.w3.org/TR/css3-values/#angles )

>From that text, I expect that also turn units are normalized to
0-360deg, because normalization is used to ensure consistent
computation of values and animation between values in different units.
This means that the computed value for any integer value in the turn
unit is 0deg, as far as I understand.

However, that requirement is a should, so implementations may just
ignore it in the computed value of a transformation function.

IMHO, the best approach here is that every property says if
normalization must be applied in the computed value or not, like now
every property says that relative lengths are computed to absolute
ones.

> Xmlizer
>
>

Giovanni

Received on Monday, 23 March 2009 17:37:40 UTC