W3C home > Mailing lists > Public > public-fx@w3.org > October to December 2011

Re: Merging CSS Shaders proposal into Filter Effects

From: Dean Jackson <dino@apple.com>
Date: Tue, 01 Nov 2011 14:26:38 -0700
Cc: "Tab Atkins Jr." <jackalmage@gmail.com>, Vincent Hardy <vhardy@adobe.com>, Patrick Dengler <patd@microsoft.com>, Dirk Schulze <vbs85@gmx.de>, "public-fx@w3.org" <public-fx@w3.org>
Message-id: <723BA831-9B0D-40CE-877A-12A44CBECD20@apple.com>
To: Chris Lilley <chris@w3.org>

On 01/11/2011, at 11:29 AM, Chris Lilley wrote:

> On Tuesday, November 1, 2011, 1:35:00 AM, Tab wrote:
> 
> TAJ> On Mon, Oct 31, 2011 at 5:29 PM, Chris Lilley <chris@w3.org> wrote:
>>> On Tuesday, November 1, 2011, 1:18:10 AM, Tab wrote:
> 
>>> TAJ> Fragment shaders can do arbitrary pixel transformations.
>>> TAJ> There is no way, even theoretically, to make pointing respond to that.
>>> TAJ>  Some existing Filters *could* so so (like feOffset), but we've
>>> TAJ> resolved to keep them consistent and have them not affect pointing.
> 
>>> I don't recall such a resolution. I do recall some discussions, which did not reach a conclusion, about whether it would make for more intuitive for users if mouse pointer picking was affected by feOffset (for example) so that what you *think* you are clicking on is what actually gets clicked.
> 
> TAJ> feOffset is used for things like drop shadows, though, which
> TAJ> definitely *shouldn't* affect pointing.
> 
> There are cases where you want it to affect positioning and cases where you don't. For drop shadow, it depends how you use it. The regular shadow behind filled text, you don't want it to, I agree.

This is a good point, and we had a similar discussion at the last CSS F2F meeting about pointer-events. The request was to have a way to toggle transparent pixels so they would not accept clicks. Of course, this isn't always want you want - UI icons with holes are typically responsive when you click on the hole. You want to define a separate region for pointer events (eg. a bitmask).

A filtered element might want to hit test on its original location (I imagine some blurs would ask for this) or on its transformed location.

Either way, I don't think CSS Shaders is adding anything new here beyond what's already possible. I also don't think this should be used as an argument to keep CSS Shaders out of the Filters spec.

But maybe we've gone far enough by email and can just talk it through on Thursday.

Dean
Received on Wednesday, 2 November 2011 02:35:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 2 November 2011 02:35:09 GMT