Proposal: upgrade bullet physics to 2.83
abstract: Threading physics on multicore gpu, and cpu systems can allow for denser physics meshes, as well as for better continious collision detection,
it also opens up the possiblity of decomposing triangle meshes into many simple bounds or many convex hull bounds, allowing for more realistic collisions and
less errors.
our occlusion is done by bullet currently as well, I think.
part 2:
Candy
if durring this stage, LOD , occlusion and instanced geomtry shaders could all be written in OpenCL and openGL, then blender game engine would have almost the same power as ue4 graphically, with the exception of a tesselation shader.
Since logic is almost bug free (as far as I can tell)
should development of enhancements be a priority now?
part 3 imposter mesh = push our instancing system into modern geometry instancing, with batched draw calls. if you have 1000 of the same item in view,
you can use 1 draw call, rather then 1000.
use = add instanced duplicates to scene
part 4= automatic dynamic lighting manager = moves x lights around,
populating light positions, allowing us to use hundreds of lights in a scene,
as most of the time, you never need more then 12 lights in a visible area.
if I were to kickstarter this, who would code it?
how much do you think it is worth?
10,000$ us?
What am I missing to make the bge a contender ?