W3C home > Mailing lists > Public > www-style@w3.org > November 2010

RE: [CSS3-UI] Problems with pointer-events draft

From: Kevin Ar18 <kevinar18@hotmail.com>
Date: Wed, 17 Nov 2010 19:36:13 -0500
Message-ID: <SNT110-W47911C981F368E83F1F84AA390@phx.gbl>
To: <tantek@cs.stanford.edu>
CC: <www-style@w3.org>

Thanks for the pointer.

Question about Issue 5:
"In SVG 1.1 the 'pointer-events' property only applies to 'graphics elements', but in css3-ui it applies to all elements. Why?"

I am really concerned about this issue.  There was a long discussion among the SVG working group, the CSS working group, and several browser developers over this very question.

The conclusion was that the svg tag should not trigger pointer events like the SVG spec states and as noted here: http://www.w3.org/2010/09/27-fx-minutes.html#item03
Originally, the plan was to include this explanation in an SVG spec, but it was concluded that it should be in the CSS specs instead.

Has something changed?


> From: tantek@cs.stanford.edu
> Date: Wed, 17 Nov 2010 15:14:46 -0800
> Subject: Re: [CSS3-UI] Problems with pointer-events draft
> To: kevinar18@hotmail.com
> CC: www-style@w3.org
> 
> Hi Kevin,
> 
> Thanks for reviewing this CSS3 UI editor's draft feature and providing
> early feedback - really appreciate it. I've collected your list in the
> issues list and will expand and follow-up there.  Feel free to add
> more there as well.
> 
> http://wiki.csswg.org/spec/css3-ui
> 
> Thanks,
> 
> Tantek
> 
> On Wed, Nov 17, 2010 at 12:47, Kevin Ar18 <kevinar18@hotmail.com> wrote:
> > In reference to the following section:
> > http://dev.w3.org/csswg/css3-ui/#pointer-events
> >
> > I figured I would polish up some test cases based on what was put in the
> > CSS3 spec.  However, there are a number of points of confusion that I want
> > to clear up.
> >
> > First, off, these are only my views on the issue... so I really need some
> > input to see if my reasoning is wrong or not.
> > Second, I apologize for the the aggressive sounding tone; I really don't
> > mean it to be that way ... and I appreciate the effort people put into
> > working out the details ... however, I just want to get to the heart of the
> > issue quickly, so I really hope the original author(s) won't feel offended
> > by it. :)
> >
> >
> > Issue 1. As far as I can tell, the spec does not address the most important
> > issue: whether the svg rectangle should trigger pointer-events.
> > There was considerable discussion over this very topic in the svg mailing
> > list as well as in various meetings.  The conclusion: it was decided that
> > the topmost/base svg tag should not trigger pointer events -- it should not
> > act like an invisible rectangular object that you can click on.
> > The relevant discussion is listed here:
> > http://www.w3.org/2010/09/27-fx-minutes.html#action04 under the section:
> > "'pointer-events' and (event) transparency"
> >
> > Sounds fine... but... I do not see that mentioned in the spec ... which
> > means is this issue still undecided?  is it just an oversight?  am I
> > mis-understanding something?
> >
> >
> > Issue 2. Why even mention all the pointer-events properties for SVG when
> > they all mean the same thing for the svg tag: "none"?
> > * According to the SVG spec and the discussions, the svg tag never triggers
> > pointer events -- no matter what the "pointer-events" property is set to.
> > (There is one exception: the "boundingBox" property from SVGTiny)
> > * Thus, "pointer-events" has no effect on the svg tag; it only affects it's
> > children. (Again, "boundingBox" is the 1 exception.)
> > * Example: auto, none, all, visible, visiblePainted, visibleFill,
> > visibleStroke, painted, fill, stroke ..... do not affect the base svg tag in
> > any way; these properties are merely inherited by it's children.  Note:
> > "auto" is the only property that is different. "auto" changes to
> > "visiblePainted".
> >
> > So, why even list many of the attributes like "visible", "visiblePainted",
> > "visibleFill", etc... when they have no effect?  Or..., does this imply that
> > the CSS3 specs actual means something different?  Do things like "visible"
> > and "visiblePainted" work differently than in SVG?  Really, that is probably
> > the most imortant thing; does listing this info in the spec imply that there
> > is something different about it?
> >
> > Issue 3. There is no way to specify that the svg tag should trigger events.
> > Based on the points 1 & 2, none of the current properties allow the svg tag
> > to trigger events.  I can forsee this being a pretty big issue when you
> > specifically want to treat the svg element like a drawing surface and
> > capture mouse events.
> >
> > SVG Tiny added another property called "boundingBox"
> > (http://www.w3.org/TR/SVGTiny12/interact.html#PointerEventsProperty)  Should
> > a property like this be added to the CSS3 specs to allow the svg tag to
> > trigger events?
> >
> 
> 
> 
> -- 
> http://tantek.com/ - I made an HTML5 tutorial! http://tantek.com/html5
 		 	   		  
Received on Thursday, 18 November 2010 00:36:48 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:34 GMT