W3C home > Mailing lists > Public > public-fx@w3.org > October to December 2012

[filter-effects] FPWD publication request

From: Dirk Schulze <dschulze@adobe.com>
Date: Wed, 3 Oct 2012 13:15:00 -0700
To: "public-fx@w3.org" <public-fx@w3.org>
Message-ID: <80F884D5-1F0E-4381-BD9D-D1758073BDD6@adobe.com>
The SVG working group[1] and the CSS working group[2] resolved to publish the first public working draft (FPWD) of the "Filter Effects 1.0" specification[3]. The specification introduces filter functions[4] and new filter primitives like custom shaders.

In the meantime members of the W3C raised concerns to the current specification. I want to address these concerns.

1) GLSL is the recommended shading language for custom shaders
The specification lists a number of potential shading languages[4] that were considered for custom shaders (including GLSL, HLSL, Adobe's Pixel Bender and Apple's CoreImage Filters). The editors strongly believe that a recommended shading language is in the interests of content creators / web authors and for the web in general. The restricted version of GLSL form WebGL was chosen because of its wide implementation across browsers like Firefox, Safari, Chrome and Opera (some still in experimental state).

2) No normative shading language (like for media sources[5]) above "RECOMMENDED" shading language
The definition of "Recommended" according to RFC 2119[6]:
"3. SHOULD   This word, or the adjective "RECOMMENDED", mean that there
   may exist valid reasons in particular circumstances to ignore a
   particular item, but the full implications must be understood and
   carefully weighed before choosing a different course."
The editors believe that a recommended shading language is in the interests of the web, since it avoids fragmentation and implementation coasts for web authors. However, if there are "valid reasons", an implementation can differ from the recommendation. Since projects like ANGLE demonstrated that GLSL can be parsed and translated, a different course is harder to defense.

3) The specification should allow more shading languages to be future proof
This particular concern seems to be very valid. GLSL might have different versioning in the future, other shading languages might be supported by
UAs as well. Filter Effects should allow, but not require support for more shading languages then GLSL. I added an issue to the specification that addresses this concern.

4) GLSL as RECOMMENDED feature makes it an essential feature under the W3C Patent Policy
The section about "Limitations on the Scope of Definition of Essential Claims"[8] explicitly excludes implementations that are not part of the normative portions of the recommendation. This is the case of GLSL.

No member officially objected to the publication on one of the working groups, even after a request to do so[9]. The raised concerns don't have enough substance to prevent the publication process. Therefore I asked Chris Lilley, the team contact of the SVG WG, to execute the resolution of the working groups.

Greetings,
Dirk

[1] http://www.w3.org/2012/08/09-svg-minutes.html#item01
[2] http://lists.w3.org/Archives/Public/www-style/2012Aug/0898.html
[3] https://dvcs.w3.org/hg/FXTF/raw-file/e2dde2a58f38/filters/index.html
[4] https://dvcs.w3.org/hg/FXTF/raw-file/e2dde2a58f38/filters/index.html#shading-language
[5] http://dvcs.w3.org/hg/html-media/raw-file/tip/media-source/media-source.html#mediasource-methods
[6] http://www.ietf.org/rfc/rfc2119.txt
[7] http://code.google.com/p/angleproject/
[8] http://www.w3.org/Consortium/Patent-Policy-20040205/#def-essential-exclusions
[9] http://lists.w3.org/Archives/Public/public-fx/2012JulSep/0097.html
Received on Wednesday, 3 October 2012 20:15:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 3 October 2012 20:15:27 GMT