[ODE] asynchronous collision detection

[Adam Rotaru]

--- Nate W <coding@natew.com> wrote:
> Seems to me that if you run the physics twice
> without checking for
> collisions, you'll get more interpenetration than
> you want, so you'll get
> awkward large forces to counter un-penetrate things.
>  And if you run
> collision checking twice without running the
> physics, the second collision
> test will yeild the exact same results as the first
> pass, because nothing
> will have moved since the first pass.
> 
> Sounds to me like you'll either get unstable
> simulations or you'll be
> wasting processor time, neither of which is
> desirable.
> 
> If there ARE advantages to doing things
> asyncronously, I'd like to know
> what they are and how it works, because it sounds
> counter-intuitive to me.

This all sounds correct to me.  One more possibility 
--
and I don't know if this is possible or makes sense --
is this:
 1 run a collision detection
 2 run a physics simulation step with the contacts
 3 run one (or more) additional simulation steps with
*no* contacts
In 3 non-penetrating bodies will move correctly, and
penetrating
bodies will also move with their current speed (either
increasing or decreasing
penetration depth).  I think it would not be unstable.

--adam



__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/