Real-Time Physics Simulation Forum

Right, this type of thing would accelerate the acute angle projection problem. But (and this is a pretty big but), it breaks down if there are other walls in the scene, as the intermediate acceleration step moves the character a long way away from the allowed region, possibly into proximity of other...

Here's a picture that might explain a little bit of what's going on with the negative slope issue (just a gimp-job, not results from a simulation or anything...): http://www.gamefight.org/iter.png This is a position space picture of a two element chain attached to the ceiling, where the x-axis repre...

bone: I agree, it very well may cause problems with the overall simulation, as it is trying to cheat the iteration process. I can think of a few cases where it should actually be all but provably impossible to end up with exactly the right outgoing momenta without actually going through each theoret...

If you do Java, Phys2d (another Box2d port/edit, I'm not sure how close it sticks to the original) implements bouncing, at least according to the Javadoc, but I'm not sure exactly how, where, or whether it's the right way or not.

http://www.cokeandcode.com/phys2d/

Yup, that (your use of the algorithm and what it was/was not intended to do) makes sense. I'm not going to drop the discussion quite yet, however, because EternlKnight's original question did seem to relate to some of the things that your algorithm didn't address, but I suspect it may give a clue as...

Okay, I looked into this a little bit more, and from what I can tell, this actually does seem to work at least to the level of physical plausibility. Here is my new linear constraint function, trimmed of a lot of irrelevant custom bookkeeping/optimization/early exit stuff: [this is Java code, BTW] /...

Right, that's an interesting approach - am I correct that when you say you limit the impulse that means that you keep summing the impulses and then constrain them to the limiting shell, so that even if you've maxed out the impulse along the normal you may still acquire a tangential component? Otherw...

I've recently run into some nasty issues of this sort (light things getting hammered through walls) in a 2d Verlet engine I've been playing with - if you're using thin enough stationary walls (in my case they are necessarily infinitesimally thin lines with no clearly defined inside or outside, so al...

Seeing as you can't load that applet (it's a general JOGL applet problem on some computers, I have no idea what causes it since I've never come across it myself, but it's not exclusively a Linux thing, it's been known to happen on OS X as well), I figured I'd run a test and post the screenshot: http...

Sorry, lost track of this thread for a while, just checked in today. Upon review I have to take back a lot of what I said earlier about this not giving a good measure when interconnections are present. I opened up a debugging level for a Verlet-based game I've been working on and altered it so that ...

Having tried something like this before, I can tell you that this method works, kind of sort of, depending on what you want to use it for. Under a Verlet scheme, you're using the position and last position as stand-ins for the velocity, so there is a linear relation between the displacement of the c...