I would rather see ODE and the internal system deprecated and slowly phased out. I like both of them, but they can’t compete with Bullet, which is more robust, more complete, much faster and a lot easier to use. I don’t think having official support for more than one physics engine makes sense in the long run. Especially since Panda does lack manpower.
This lack is one of the reasons I started wrapping my head around C++, but some time will pass before I’m of any use in this matter…
The way I understand it, it was a student project, so its main purpose was not necessarily for people to actually use it. With PandaAI being limited to a very simplistic 2-dimensional graph, it just doesn’t make much sense outside simple tech demos.
Recast would be a much more reasonable option for actual use. It lacks documentation and so on, but it has incredible power. It’s like the Bullet of AI. But we all know that. I’d like to integrate Recast and Detour as a part of my editor project, but I’m not sure how it will work.
As far as inconsistencies between D3D/OGL or GLSL/Cg, I can’t say I really care about them that much. I would prefer work to focus on OpenGL, because it’s multiplatform, and Cg, because it already works well, rather than trying to bring D3D up to speed – because the way I understand it, D3D backend is simply lacking, right? I’m probably biased because I’m a Linux user, but if you have an engine that’s supported on 3 platforms, why waste time on something that only works on one of them?
Speaking about features, I wonder what happened to the project someone was working to add support for 32bpc float textures. Was this effort of any success, or did it just die?
Also, I think something like a per-project donation system would be useful. Something like out own little kickstarter where people could support the specific community projects and efforts. On many occasions users wrote that they would pay to have something done or added. Why not allow them to? Obviously, someone would have to set the whole thing up, which is where we’re back to the lack of manpower ;D. Plus, with most of us being C++ illiterate, we wouldn’t be able to add requested stuff in a proper way and reasonable time even if we got payed a million $.
I don’t want it to sound like I’m suggesting money solves problems (actually, I think it creates them, but that’s a different topic), but it certainly helps get people to sit on their asses for 8 hours a day working on Panda .