Ran into a bit of a problem, though not an unexpected one. In my current program, I’ve got a rudimentary collision detection in place, but there are a number of situations where things don’t work as expected. For the sake of simplification during development, all the objects involved are spheres. When the velocity of the colliding objects is greater than twice the radius, there’s a chance that two objects will pass through each other completely. If the velocity is greater than e, there’s the chance that the collision will resolve from the opposite direction, accelerating both particles involved.
The quick hacky fix is not letting the velocity go above the radius of the sphere. Another is to up the refresh rate. But fundamentally, the essential structure of the collision engine has to change. Whether that means a more in depth exploration of collision, or speeding up my transition to full Java/importing a preexisting physics engine, I’m not sure yet.
Rather than get frustrated or discouraged, I’m trying to sidestep the issue for now, and develop more of my mini ideas. There’s lots of work to be done figuring out things like direction, art, and fun.