Hi,
no, i dont have any optimizations, its pretty straight forward.
the article where this is all based from has an approximation of the square root which seems really good but i never got to make it work, i dont know why but i got weird results...
i've tried other sqrt approximations and i got the exact same results, it was a slower. i think it because modern processors already come optimized for square roots.
im working a lot on improving this, i'm adding polygon to polygon collision and i really want to implement some sort of broad phase algorithm to speed things up.

sounds great! I will make mine really simple since i only made it to learn a little more about how those things work. The fastest way of speeding up collisions for now is to sort all particles depending on their x position and then using a smart loop that breaks; at a certain point. That only seems to be useful if all particles collide though!

yea i saw an example from Zach Lieberman where he had the particles sorted that way.
i've also downloaded some articles explaining the prune and sort algorithm, if its not too hardcore to implement i might go that way, otherwise i'll use Zach's algorithm :)

hehe, I just read some things about it. it seems pretty similar to the sorting i described only that it's bounding box based and a little more complex if I am not totally mistaken?

from the little i've read about it, yes thats pretty much it :)

oh yeah!!!

Hi,
this is a little different than Box2D and it suits just fine what i needed it for: animating particles with forces and constraints.
At the time i though Box2D would be overkill and im also really interested in generative animation so it was also an excuse for me to learn a lot more about the subject.
:)

Makes sense, thanks for sharing.