As I was thinking of ways to separate my destructible Body hack for Phys2D into something more robust, efficient, and less hacky, I thought that maybe Constructive Solid Geometry would be a good choice. I would add a new type of Body to Phys2D that would have a binary tree of operations (Body’s added or subtracted). The actual used collision would be the union of the added shapes minus the union of the subtracted shapes. This would allow for much more intelligent destructible terrain without removing any of Phys2D’s robustness or efficiency. It also allows absolutely any type of shape to be constructed with minimal fuss and a simplistic view of geometry.
I’m pretty sure I have the know-how to do this, but before I delve in… any reason why this might be a silly idea?