Advanced Collision Rebound From Curved Surfaces

Hey. First post, and I’ve got something of a doozy to ask. I’m trying to get do a couple things for a project of mine, and I think I may have, er, overstepped the limits of my skills. What I’m trying to do is make an object react correctly to collisions with curved surfaces.

Things are not going so well, and while I have a few ideas left untried, it seems wise to ask around for other thoughts before I try to implement things that may not work at all.

My thoughts on the subject so far run as follows:
Initially, I tried using simple movieclip objects overlapping each other to provide borders for the spaceship, such as basic circles, squares, and concave areas. With a little stretching and rotating, these could be made to form the levels just fine, but the problem in every case was finding the tangent line to the point of impact on the shape. This fault led me to frustration and despair. Then I decided to come here before scrapping it completely and moving on to my next plan…
The equation of the tangent line at point of impact is, by definition, the value of that function’s derivitive at that point. So the thought occurs to me: What if I could make Flash into, in essence, a graphing calculator designed to display parametric equations? I would have to make the derivitives by hand, but that wouldn’t be hard… Possibly an new object, Equation, might help, that would contain the equation, it’s derivitive, and the domain of *t[/t]. There might be a little trouble determining the exact point of impact, but in theory it should work if I can get Flash to view the graph as a single object.

I probably made a mess of my explanations, heh… If you can understand my dilemma, any insights would be helpful.

Er, bit101? Enlighten me, please, I’m new to this kinda stuff. :stuck_out_tongue: Despite the complexity of my question, heh, I’m rather amatuerish at this stuff. Thanks for the prompt reply though. =)

Sorry, I thought it was curved, but it was a straight line. So I deleted my post because I thought you woudln’t find it useful. But anyway. You can view it here: www.bit-101.com go to Laboratory, and in the calendar go to 03 OCT 09.

If the line’s slope isn’t 0 or infinity, it’s the same kind of problem, and I don’t think bouncing off of horizontal and vertical lines is the subject of much study. :smiley: Thanks a billion, I’ll check it out right now.

…need more caffiene…