Re: [css3-3d-transforms] Formal definition of 'backface-visibility'

Just as a follow-up to this, I have since switched Gecko to the former approach since it gives better results.

https://bugzilla.mozilla.org/show_bug.cgi?id=693520

- Matt


On 26/07/2011, at 5:18 PM, L. David Baron wrote:

>  http://dev.w3.org/csswg/css3-3d-transforms/#backface-visibility-property
>  doesn't have any formal definition of when the "back" of something
>  is visible.
>
>  For example, some obvious possible definitions would be:
>  * if the element's transform involves projecting the vector [0 0 1]
>    to a result with a negative Z component
>  * if the element's cross product of the transform of [1 0 0] and
>    the transform of [0 1 0] gives a result with a negative Z
>    component
>  These two give different results in some cases (I think when the
>  transform matrix has a negative determinant).

Yes, we need to define this. I think it's even a little more complex than this, because it involves the parent transform as well.

I'll try to come up with something and send it here and to the Mozilla bug.

Dean

>
>  Matt Woodrow implemented the latter in Gecko in
>  https://bugzilla.mozilla.org/show_bug.cgi?id=505115  ; I'm not sure
>  if that was for compatibility with WebKit.
>
>  -David
>
>  -- 
>  ?   L. David Baronhttp://dbaron.org/    ?
>  ?   Mozilla Corporationhttp://www.mozilla.com/    ?
>