- From: Dean Jackson <dino@apple.com>
- Date: Wed, 05 Oct 2011 10:27:00 +1100
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- Cc: "Gregg Tavares (wrk)" <gman@google.com>, robert@ocallahan.org, James Robinson <jamesr@google.com>, www-style list <www-style@w3.org>
On 05/10/2011, at 10:18 AM, Tab Atkins Jr. wrote: > On Tue, Oct 4, 2011 at 4:05 PM, Gregg Tavares (wrk) <gman@google.com> wrote: >> On Tue, Oct 4, 2011 at 3:35 PM, Robert O'Callahan <robert@ocallahan.org> >> wrote: >>> On Wed, Oct 5, 2011 at 9:13 AM, Dean Jackson <dino@apple.com> wrote: >>>> >>>> As for filters, I was really hoping that we'd be able to specify them as >>>> purely a 2d effect - as in whatever is composited in the end is a quad, not >>>> a mesh, even if the filter itself uses a mesh. >>> >>> I totally agree. The result of a "custom" filter primitive is flattened to >>> an image. Nothing else makes sense, even within a filter chain >> >> That's fine in all you want to achieve is filtering effects but that doesn't >> fit with the examples shown by Adobe. They're morphing the content in 3d. >> http://www.adobe.com/devnet/html5/articles/css-shaders.html#1 >> In one case they're morphing something to look like a newspaper. It unfolds >> but is still interactive once it unfolds. >> In another case they're morphing a twitter feed a kind of wavy shape and >> they are interacting with it while it's in this wavy shape. I don't see why >> the user would think it's not 100% interactive. >> So it seems like for this proposal to be generally useful that content needs >> to able to be interactive (ie, the user needs to able to click on stuff just >> like they can with 3d css). Otherwise the number of use cases for it seems >> exceedingly small. > > Note that vertex shaders (what, I think, were used to achieve the > effects that Greg mentioned) are reversible, though it requires the > shader to be run on the CPU (I think), so you can still handle mouse > interaction. Pixel shaders aren't, but that's not something fixable. The problem is that it might be computationally expensive to do this mapping each time the pointer moves, and also is a fairly high burden on some implementations (write your own GLSL engine). Dean
Received on Tuesday, 4 October 2011 23:27:39 UTC