[css3-2d-transforms] 2D Transforms feedback

[Sylvain Galineau]

Some feedback on the current public draft [1].

1. The Introduction states that a transformed element "...acts as a containing block for fixed positioned descendants". It notes that more detail will be needed in the future to explain while that should be true. I'd love to hear more.

2. As stated into the red-color part of the Intro, we need to figure out how handle fixed backgrounds. Currently, the tentative proposal is that fixed backgrounds "ignore the transform completely, since - even transformed - the object should be acting as "porthole" through which the fixed background can be viewed in its original form." One could however imagine cases where the author does want the transform function to affect the element's 'part' of the background. If, after all, transforms do not affect layout and effectively occur after the element has been positioned, the background has been painted and could be transformed as well. This seems to be the way Gecko and WebKit implement this now.

3. What are common use-cases for the Point interface and Window object extensions defined in section 7 ? Is this the way one would converts mouse event coordinates into a transformed node's coordinate space for instance ?

4. Also regarding section 7, interested in answers to  Øyvind Stenhaug's earlier question here: http://lists.w3.org/Archives/Public/www-style/2009Sep/0120.html

5. Valid input ranges for transform functions should be specified e.g. can scale parameters be negative ? Safari 4 seems to ignore the sign if the input is negative.

[1] http://www.w3.org/TR/css3-2d-transforms/
[2] http://www.w3.org/TR/css3-2d-transforms/#dom-interfaces

[Dean Jackson-7]

Hi Sylvain,

Answering parts of your feedback here.

On 01/10/2009, at 6:43 AM, Sylvain Galineau wrote:

> 3. What are common use-cases for the Point interface and Window  
> object extensions defined in section 7 ? Is this the way one would  
> converts mouse event coordinates into a transformed node's  
> coordinate space for instance ?

Correct. As Øyvind notes elsewhere, these methods are a little weird.  
They probably should be in CSSOM View.

> 4. Also regarding section 7, interested in answers to  Øyvind  
> Stenhaug's earlier question here: http://lists.w3.org/Archives/Public/www-style/2009Sep/0120.html

I'll answer that one directly, but the summary would be that yes, we  
know they are deprecated. But there hasn't been any proposal for a  
replacement - the CSS OM specification seems stalled. Do you have any  
suggestions?

> 5. Valid input ranges for transform functions should be specified  
> e.g. can scale parameters be negative ?

Good point. Negative values should be allowed for scale.

> Safari 4 seems to ignore the sign if the input is negative.

This sounds like a bug. I know it works in some cases though, since  
I've used it in public examples.

Dean

>
> [1] http://www.w3.org/TR/css3-2d-transforms/
> [2] http://www.w3.org/TR/css3-2d-transforms/#dom-interfaces
>
>
>