Re: Updated CSS shaders proposal

On Oct 20, 2011, at 5:59 PM, noam.rosenthal@nokia.com wrote:

> 
> From: www-style-request@w3.org [www-style-request@w3.org] on behalf of ext Gregg Tavares (wrk) [gman@google.com]
> Sent: Thursday, October 20, 2011 5:41 PM
> To: Vincent Hardy
> Cc: Chris Marrin; www-style list; public-fx@w3.org; SVG WG
> Subject: Re: Updated CSS shaders proposal
> 
> On the interactivity issue, issue #3
> 
> > In other words, elements that have CSS shaders applied direclty or indirectly will not receive mouse events nor touch events at a minimum.  Similarly 'hover' will no longer work.
> That sounds unnecessarily harsh... we don't have to turn them off, but rather just defer that issue of vertex-shader-assisted interactivity for later and currently keep the "default" behavior where touch/mouse event coordinates are relative to the element and not to the vertex shader positioning. This way, for example, a square that contains something like a water effect can still receive mouse events, without the web developer having to worry about which actual wave the user has pressed on.

I agree, in fact in many cases this is the preferred behavior. It might seem cool to have an animated ripple where you can click on a rippling check box, but is that really practical? Do you want to play "hit the moving target"? A better approach would be for the element to be rippling when the mouse is outside and have it settle down when you mouse over. 

I know there are cases where you want a static effect to deform an element and have clicks accurately reflected. But I'm comfortable without that capability in the first release. We should specify that the CSS filter is a visual effect which does not affect interaction. Later on we can add a filter-interaction property to allow you to turn on the feature of transforming clicks to the deformed image. I don't know how we go about doing that. But I'm happy to have it be something we defer till later...

-----
~Chris
cmarrin@apple.com