[ODE] box-box collisions

[Gary R. Van Sickle]

> > This sounds like it doesn't bode well for any sort of "collision
> > anti-aliasing" schemes, i.e. preventing a fast-moving small object
> > from going right through a wall.  Are you saying here that even if a
> > suitable collision point was found, the fact that it was by definition
> > outside the collision volume (because there is none) would prevent it
> > from being properly taken into account?
>
> i'm not quite sure i understand the question - what i was saying is that
> more stable dynamics results when the contact point is inside the
> intersection volume. if your objects don't collide, there is no
> intersection volume and thus ODE won't generate a contact point.

As I understand it, you actually mean "If your objects don't intersect at any
timestep..." in that last sentence, right?

> if you
> generate one anyway for fast moving objects (e.g. projecting the object
> along its path) the contact constraint will still act in the same way,
> but finding a contact position that will give you consistent stability
> is going to be a problem.

I guess what I'm wondering is, say you shoot an ODE bullet at an ODE wall.  At
one timestep, the bullet is on one side of the wall, at the next it's on the
other side, in neither step do their respective geometries overlap.  Are you
saying that it isn't possible to generate a contact point that will ensure
stability in such a case?

--
Gary R. Van Sickle
Brewer.  Patriot.