"Slime" wrote:
> Yes. Let's say I have a system with two balls. Imagine this now: One ball is
> at (0,0) and is heading towards (1,1). The other ball is at (1,0) and
> heading towards (0,1). Assume their radii are extremely close to zero.
>
> They're going to bounce off each other at time=0.5 (assuming time=1 is the
> end of the iteration, when they reach their destinations).
>
> I detect the collision at t=.5. I then figure out their new velocities. (In
> this case, their velocities will simply be swapped, I believe. Or something
> like that. Let's say for simplicity that they swap their velocities.)
>
> Then I place the balls at (their current starting position) + (.5*(their old
> velocity)) - (.5*(their new velocity))
>
> Their new position is at (1,0) and (0,0), respectively. In this example,
> they happen to have swapped starting positions.
>
> Now that I've taken care of that intersection, I ignore any intersections
> that occur at t <= 0.5. That way I won't detect their collision again. And
> at the end of the iteration, I just add each balls' velocity to it's
> position, so the first ball ends up at (0,1) and the second at (1,1).
>
> Did that make sense?

Sure it did. It sounds very interesting.


> Before long, I'll work on the code and make a new example for you. But not
> now... Warcraft III takes priority =) (two levels left! I think.)

Hehe, must be real fun... :-)

--
Jonathan.

Post a reply to this message