Mark Stock's work intrigued me. The sheer details he went into to tweak the behavior and output was fascinating. I was especially impressed by the renderings he had made of paths traced by objects based on mathematical properties. I wanted to do something similar but the road block was not being able to capture particle paths and build a 3D model with it. After a discussion with David Tracy a couple of days back about this, I got a suggestion that I could try generating the point clouds by a script in Python and feed it into Rhino 3D to make the curves. So, I dived in. What I went through in the last two days was an amazing journey to figure out how it could be done. I wanted to start out with a simple mover object. Since time parameter does not exist, I simulated that with a variable and captured the co-ordinates in an array list instead of drawing the particle (the way it is done in Processing). The trickiest part was to figure out vector math in Rhinoscript. It treats simple variable list with (x,y,z) as point3d and vector3d objects. I finally got the mover and attractor working last evening. Here are a few attempts with the same. As the title of the post says, it is a "work in progress", I am planning to take this further, with a bit more complex algorithms with mutual attractions implemented. These are a few forms I made with basic attraction around the origin with a wind force that creates the randomness.
It has been modeled in rhino by python scripts and rendered in Keyshot.
Below are some screenshots from rhino.