programming

physics poking

man. i sure love fiddling with my contact constraint. i have spent an incredible number of hours just experimenting with this and that idea (or getting up in the middle of the morning to implement something i has been thinking about in bed for 2 hours (like now) ). the more i fiddle, the more intuitive a grasp a get that allows me to push it a bit farther.

my finding from playing around today was that some of the code that helps prevent jittering was also messing with the constraints ability to handle high mass ratios.

when i test my physics, i never have sleeping on and i always use smaller box sizes. my favorite size box is a 20cm cube. that is a nice size to work with and shows many issues that a 1m box just won't show. i really should push my test size down even smaller. i chose 20cm because awhile back i was trying to get things running at 25hz (so 1 tick per frame on a pal console), objects smaller than 20cm tend to penetrate too much in a single frame. tunneling is an entirely different problem to solve and i did not want to solve the two problems at once. now i work at 60hz and i probably could go down to 10cm, no problem.

anyway, after fiddling some more i managed to get 20 times mass ratios working pretty nicely. without 0 mass tricks either. for some reason i have in my head that i want 100 times ratios working well. arbitrary goals are always fun.