Which way is the correct way?

1) Add to the physics engine with the static version of the mesh (everything collapsed together) to make a btConvexTriangleMeshShape hull proxy

2) Add to the physics engine with the frame hierarchy structure to make the same thing.

Any ideas?

Thanks

Jack

For static objects, number 1.

For objects that can move, neither. Don't use triangle meshes in this case.

For moving objects, what can be used instead?

Thanks

Jack

The specifics vary by engine, but you're most likely looking for rigid bodies of some kind. Spheres, boxes, and capsules are the most common geometrical forms.

Hi ApochPiQ,

I checked on one of the forklift samples found in one well-known library called bullet.

In that sample, I see that the fork of the forklift can actually come really close to the body of the forklift,

In other words, there is no bounding volume above the fork....

Are boxes, capsules still good for those cases?

Thanks

Jack

Shapes that need to move together are usually connected using a constraint system. I only glanced at the Bullet physics demos, but it seems like the forklift is a set of boxes connected via different types of constraint.