Re: Implementing <use> without cloning

Hello Olaf,

--Original Message--:
>Another interesting question is, what is expected, 
>if the display property of a use is animated - 
>does this mean, that the appearence of the 
>embedded XHTML changes?

Yes.

>Because XHTML and CSS themselves do not really define, 
>what is animatable or not, SMIL animation is not directly
>applicable for XHTML+CSS (apart from the specific 
>XHTML+SMIL note from the SMIL group), however if an 
>animated property is inherited, the animation takes place 
>in the well defined SVG environment.

Correct.

>As far as I can remember, for example Opera did or does 
>something like a raster image of the XHTML fragment - could 
>be difficult to apply animated properties to such an image
>as it is difficult to use forms and similar interactive content
>from XHTML pretty relevant for SVG authors - even
>more than inherited, animated styling, because the functionality
>is not available in SVG at all in contrast to the presentation
>of simple text like 'Hello Kitty' ;o)
>
>Note, that the appearence of the XHTML fragment may
>change as well, if width and height of the foreignObject
>are changed with animation, therefore to get
>something useful, maybe one has to clone these things
>anyway separately for any use reference.

Semantically, cloning is the model. However, the choice whether
to actually clone is an implementation decision. It's still
possible to optimize using a raster as long as it's invalidated
in the presence of an active animation and the result is correct.

The method by which acceptable performance is achieved is an
implementation detail.

Alex

>Olaf
>
>
>

Received on Thursday, 15 October 2009 10:17:11 UTC