ofxODE is a simple backend for simulating systems of differential equations, and thought someone might be interested in it. Equations can be concisely specified, and ofxODE will keep track of the variables as you step forward in time. Parameters, values, and equations can be modified on the fly.
Attached is an image of the Lorenz Attractor ( http://en.wikipedia.org/wiki/Lorenz-attractor ), from the example app. You can get pretty cool dynamics from simple sets of equations. However, most existing simulation packages are aimed repeated trials of a given system, rather then any sort of dynamic interaction. ofxODE should make it simpler to interact with simulations…
thanks! Everything’s done with Euler’s method ( http://en.wikipedia.org/wiki/Euler-integration ), which is supposed to have a few problems numerical stability – I haven’t run into any yet, but YMMV. The delta-t for the example simulation is relatively long, in that the line segments that compose the curve are visible, but that parameter is easily changed. Smaller values will yield more precise results.
i see. i just forked it, maybe i can add the midpoint method some time soon.
your repo structure is quite strange, i got to admit. for example, your addon source code is in the repo twice (including the example), and you included compiled libs. a nice .gitignore file would be useful.
i’ll try to take a look tonight, but can’t promise anything, I was in over my head before with this.
“nan” would be “not a number”, so maybe a problem with the starting numbers?