Thanks, J-S
Thanks a lot for your thoughts. I understand your concerns about abusing the StateAttribute mechanisms and wanting to better integrate VirtualProgram into state handling. If it can be done with minimal changes to osg::State, and without affecting existing functionality, I would encourage you to try to make it cleaner and submit the changes with your VirtualProgram example.
I will look into this but I don't have much spare time now so it may take bit longer than desired. Neverthless, in the weekend I will be looking at osg::State code to think about possible modifications. Before making actual changes and submissions I would like to hear what Robert and OSG People think about VirtualProgram programming interface/usage pattern. If its interface is generally accepted I will pursue this further. There are probably few topics that may need some extra discussion when I will have the idea what modifications I will be going to make...
I'll let Robert discuss possible ways forward for the VirtualProgram when he gets better, but honestly, I'd like to be able to use this as it is right now in our software. :-) It solves most of the problems we have about using different shaders and shader permutations.
The only thing that's keeping me from using it is that it might change a lot between now and when it will be integrated into OSG. But it's a very powerful tool. I really hope others see the value as well as I do.
Our engine become heavily dependent on VirtualProgram throughout the passed year. So even if it won't fit into OSG framework we will still use and develop it. I may post updates every time we do a modification.... ;-) Just kidding I would prefer to see VP or some equivalent solution adopted in OSG.
Cheers, Wojtek _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
