I think Persian_Gamer is right. Even if you continue to apply force to the ball on the peak of it''s path, if the resulting velocity is too small it''ll be set to zero. And so the ball will stop when it reaches the peak.

You''d have to make sure that the resulting velocity when applying gravity for example is always greater then the cutoff (heavy object -> high inertia -> might not start moving).

The other problem of the ball not going through the ground can be solved by making your physics engine a bit more realistic. Simply make the ground apply a negative force to the object equal to the force of the object pushing ''into'' the ground.
Of course the amount of force the ground can apply to an object (in reality) is limited, and at some point the object will start digging into the ground (or if it''s a floor inside a building break through it). But for a game engine you could simply make it so the ground can withstand any object (= allow it to counteract any given force).

when you find a contact, calculate the force along the plane of contact. if the amount is < to a threshold, you can consider the object as resting.

Fnorm = F.Magnitude() - (F * N)

if (V.Magnitude() > 0.01f and Fnorm < 0.1f)
resting = false;
else
resting = true;

or something Haven''t implemented resting objects yet. However, that algo will not work nicely if an object is overhanging above another object. It''s quite complicated if you want to do it properly. What if the supporting object dissapear? The supported object should fall, but if it is resting, it won''t get updated, so it will stay in mid-air.

you still need to check for the velocity along the plane, and the plane normal separately.




resting_contact = true;

v_norm = V * N
v_plane = (V - N * v_norm).Magnitude()

if (v_norm > 0.01f || v_plane > 0.01f)
resting_contact = false;



f_norm = F * N
f_plane = (F - N * f_norm).Magnitude()

if (f_norm > 0.01f || f_plane > 0.01f)
resting_contact = false;

[edited by - oliii on August 13, 2003 6:54:43 PM]