On Wed, Mar 11, 2015 at 12:52 AM, Kenneth Russell <firstname.lastname@example.org> wrote:
This sounds both difficult and problematic.
I don't understand that. To me it seems like it'd be a rather easy change because all you'd do is disable some intentional limitations on what GLSL you can feed to the underlying implementation. Of course you'd have to know which implementations are good for it, but that work is already being done for WebGL2 anyway, and when in doubt, you don't offer the extension.
I'd rather focus on getting WebGL 2.0 out the door.
By all means focus on that. I bring it up because WebGL 2.0 will come with a large number of caveats for hardware/drivers/platforms which will result in a very low activation rate for the first years of its existence. Offering WebGL 2 features in WebGL 1 is important for two reasons:
It eases the migration from WebGL 1 to WebGL 2
It allows some of the WebGL 2 features to be used, where otherwise WebGL as a whole would not be available (because of aforementioned caveats, or simply because a vendor hasn't implemented it yet).
I believe it's important not to loose focus on the aspect that WebGL 1 will be with us long after WebGL 2 is stable and actively used. WebGL 1 and 2 aren't like CSS2 and CSS3.