Hello all, I'm just starting out with Bullet (jbullet actually, but for the purposes of these questions I doubt it matters) and I have a few easy questions that I haven't found well explained in the docs.
You know how more or less every wikipedia article starts with "[Term] is a [simple, high level explanation of what term is]."? That's what I feel like I'm missing in the bullet docs. For example, http://www.continuousphysics.com/Bullet ... bject.html mentions one thing a GhostObject can do, what the default for that action is, and three things it is useful for, but it doesn't address what a GhostObject is.
My immediate task is trying to add collision detection to a robot simulator. I've already got the robot modeled, displayed and controlled, I just need to find out when it collides with itself. Once I have that working I'll start working on dynamics. If my questions seem to be headed in the wrong direction given my goal please correct my path!
What is a GhostObject? When should I consider using one?
What is a CollisionWorld? At the high level, what does the flow look like to go from known geometric relations of the objects in the scene graph to whether or not there are any collisions going on? Presumably I start by calling addCollisionObject a bunch of times, how should I tell the collision detection algorithm to run, and how should I get the results back?
How does raycasting play into collision detection? Should I be telling the collision detection system about the movements of the parts in addition to (or instead of?) the location of the parts?
Even if you can only answer one of my questions I'd love to hear from you!
Thanks for the help!