[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] WEBKIT_ extensions

On Sun, Nov 6, 2011 at 11:23 PM, Mark Callow <callow_mark@hicorp.co.jp> wrote:
KHRONOS, or the more likely KHR, implies wider approval than just a single working group and would have to be discussed within the Khronos Group. I suspect the answer will be that WebGL only extensions should not use KHR.

It's odd if "Khronos" isn't appropriate for "Khronos ratified extensions".

The naming isn't really the issue that came up above, though. The issue is that the use of vendor prefixing for web APIs differs significantly from prefixing for OpenGL APIs. The use of them in WebGL, which is both of these, isn't yet well-defined.

Web APIs are normally prefixed in each implementation separately, eg. mozRequestAnimationFrame, while the API is being stabilized. Once an API is fixed and implementations catch up, they remove the prefix. This prevents pages from becoming dependant on earlier versions of the spec under its final name, so the API's development isn't limited by backwards-compatibility before it's even finished. Each vendor uses its own prefix, so any dependencies caused by this are localized to that browser, and the presence of a prefix is a signal that the API isn't yet finalized.

OpenGL extension prefixes don't do any of that; prefixing is used differently, for different reasons. WebGL extensions seem more like Web APIs than OpenGL APIs.

Glenn Maynard