I am currently thinking about writing an article on collision physics on Gamedev.net, and I'd like to get some opinions from this forum. Obviously, judging from the lack of articles on the subject, achieving realistic collisions is not an easy task. There are many ressources pertaining to collision detection, but very few concerning what to do after a collision has been detected. I do not claim to know the complete answer to this problem, but I have been working on a partial solution for a few months now, and I'd like to share it with whoever is interested. This is why I would like to know how many (if any) of you would be interested in an article on the subject. You can (and should) download my demo here (D3Cube.zip --- ~350K) to take a look at its capabilities. Although my engine models friction and springs to some extent, the focus is really on collisions of objects, and the physics behind it is quite neat. My engine supports - Preservation of linear momentum - Conservation of energy in elastic collisions - Totally inelastic collisions About half the scenes from the demo show decent examples of working physic, while others show the engine's limits. I have also included some of the tests from Jeff Lander's article on commercial physic engines (in the demo, press Ctrl+F1 - F5), even though my model fails miserably most of the more advanced tests. Still, it doesn't matter that much; my article would not focus on implementation details, but rather in solving the basic equations for elastic and inelastic collisions, which is something I have not seen in any article posted here. This allows for linear and angular momentum to be treated realistically in simple collisions. A lot of problems from my demo come from the attempt to treat idle objects with the same equations as colliding ones. For each frame, there is a new collision between the floor and practically every object. This is not very efficient for real-time games, and a simpler model for collisions with floors and walls would probably yield nicer results anyway. I am interested in any comments on my demo: good, bad, insults --- I just want feedback. Cedric Edited by - cedricl on March 28, 2002 10:13:27 PM

Sounds like a good idea. I''d ask that you make sure you have enough detail if you aren''t focused on implementation. Math based articles need to be exact to the point of obvious.


Good luck!

If all that matters is what you get in the end, why go through life?
~Michael Sikora