pouët.net

Yo. I got this idea back in school but never bothered to follow through; I wonder if anyone else has had the same idea and actually done it.

http://en.wikipedia.org/wiki/Z-transform

With z-transform, one could (as far as I can tell, anyway) transform physics iterations into linear functions, so that you could, for example, make a particle system in such a way that you can ask the particles' positions based on time, and get the result directly without doing tons of iterations.

Possible uses would be potentially reduced cpu usage, extremely slowed down/reversed particle system replay without having to store craploads of data, etc.

Looks like someone doesn't know how to calculate an integral in the first place... :)

And anyway, if you have particle collisions this just doesn't work.

Of course, and I can think of many other things you could potentially do with particles that would not work, like particles spawning particles, etc.

standard Newtonian physics, no need for any z-transforms. It has been done a bazillion times already.

sol_hsa:

Not really related but there are fluid solvers that operate mostly in the frequency domain.

http://www.dgp.toronto.edu/people/stam/reality/Research/pdf/jgt01.pdf
http://www.dgp.toronto.edu/~stam/reality/Research/pub.html

Not sure how that relate though ;)