HI Sukender, On Mon, Feb 23, 2009 at 9:38 AM, Sukender <[email protected]> wrote: > What Paul points out about GL 3.1 raises a simple question: Will OSG 3 be... > 1. Based on GL 3.0? > 2. Based on GL 3.1? > 3. API agnostic, but will wait GL 3.0? > 4. API agnostic, but will wait GL 3.1? > 5. other? > > Going directly to an API agnostic OSG seems as risky as interesting. And > IMHO, waiting for GL 3.1 seems very interesting too, as we may see the 3D > future's a bit clearer. However, if GL 3.1 does not come shortly after 3.0, > we may loose too much time. Any rough idea on the delay before GL 3.1 specs > are released?
First I have to say that the OSG 2.8 already has almost all of the hardware support that will be available in OpenGL 3, thanks to the OpenGL 2.x extensions that we already support. The whole idea of OpenGL 3.1 being new clean API isn't actually that critical as we already have implemented the scene graph, it's not as if we are about to write a new game from a completely clean slate, so the value in the new trimmed API is not so critical for us, OpenGL 2.x + extensions actually does pretty well. This fact puts us in a position of strength, we can move to supporting GL 3 when we ready, and the similarity between GL 3 and GL 2 + extensions actually means a lot of high level solutions will work across both as well. Taken in this context GL 3 isn't the big leap in itself. The big leap is embracing more flexible composition of state, this is in essence what the next gen scene graph will have to embrace. My current thought on the overall situation for the future of the OSG is that we need to make composition of state better across OpenGL 2, OpenGL 3, as well OpenGL ES and OpenCL + C++ rendering backends. In this context the GL3 is actually a relatively minor part of jigsaw, and rather than OSG 2.x being an OpenGL 2.x scene graph, and OSG 3.x being an OpenGL 3.x scene graph, OSG 3.x becomes a scene graph + flexible rendering backends that are associated with different types of graphics context that we create. If OSG 3.x isn't able to make this leap then I think it would fall rather short of the opportunities available in widening the OSG's platform capabilities. Robert. _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
