Re: Updated CSS shaders proposal

From: Chris Marrin <cmarrin@apple.com<mailto:cmarrin@apple.com>>
Date: Thu, 20 Oct 2011 10:43:02 -0700
To: Adobe Systems <vhardy@adobe.com<mailto:vhardy@adobe.com>>
Cc: www-style list <www-style@w3.org<mailto:www-style@w3.org>>, "public-fx@w3.org<mailto:public-fx@w3.org>" <public-fx@w3.org<mailto:public-fx@w3.org>>, SVG WG <public-svg-wg@w3.org<mailto:public-svg-wg@w3.org>>
Subject: Re: Updated CSS shaders proposal


On Oct 19, 2011, at 9:23 AM, Vincent Hardy wrote:

Hello,
I have updated the CSS shaders proposal:
https://dvcs.w3.org/hg/FXTF/raw-file/tip/custom/index.html (you may need to force a reload to see the Oct. 19th version).
the changelog is at:
https://dvcs.w3.org/hg/FXTF/raw-file/tip/custom/index.html#october-19-2011
This accounts for some of the feedback received so far. Other feedback is captured at issues (see http://www.w3.org/Graphics/fx/wiki/Custom_Filters#Issues_List).
This also adds a way to control whether or not the triangles in the shader mesh are 'attached' or 'detached', to enable a wider array of effect.

Thanks for the update. Great progress. A couple of comments:

- Make sure you consistently use the term "fragment shader" rather than "pixel shader".

- I'm concerned that you say in section 6.2, "This document recommends the adoption on GLSL ES for alignment with shaders in WebGL". I think it would be more clear to say something like "This document recommends the adoption of the subset of GLSL ES defined in the WebGL 1.0 specification." I know that gives it even more of a WebGL slant and this may be my prejudice talking. But that seems like the most clear way to state the recommendation.

>> I have made these changes, they'll be in the next rev.

- Regarding Note 12: I think the spec should use the mime-type of the url() or <script> as the type identifier. I also think it should require WebGL support and define the mime-types for WebGL vertex and fragment shaders. This still allows implementations to support other shading languages using platform specific mime-types. The reason I'm pushing so hard for WebGL (other than the obvious reasons) is that I really want this to be a consistent feature on all browsers, and WebGL is the only web standard for shaders available.

>> I added a note about using the url/script mime type. I did not add anything about the WebGL shader mime types because I cannot find this in the WebGL spec. Do you have a pointer?

Thanks,
Vincent.