Inertia of Cone Geometry seems wrong

[maheshkurmi]

Oimophysics uses Inertia about center of shape at (0,H/2,0) assuming origin at center of base of shape.

While calculating Inertia of body, parallel axis theorem is used but I think cone geometry should be treated differently as its center of mass lies at (0,H/4,0) rather than (0,H/2,0).

Correct me If I misunderstood your method of calculating inertia tensor.

[saharan]

I think you are right, but currently OimoPhysics doesn't have a flexible center of mass system and each shape's center of mass is fixed to its local origin. Adjusting cone's center of mass to physically accurate position at this point makes it less intuitive to add cones to a rigid body at an intended position (since the origin shifts from the geometrical center).

So I leave this as intended, but want to fix when a more flexible center-of-mass system is implemented (like Box2D's one).

[maheshkurmi]

I have ported OimoPhysics engine to java (https://github.com/maheshkurmi/OimoPhysics_Java).
I am considering it over Bullet jni wrapper since I can easily modify it as per my need.

Planned updates

1. Re-organisation of callbacks and structure to match with box2D
2. Masstype (as in box2d) to allow customizing center of mass location
3. Heightfields
4. Particle System

Can I use the modified engine in commercial application?