Right now I'm playing around with DynGen, a routine which uses a hydrodynamical approach to simulate a dusty torus and a planet orbiting within it. It uses laroche limits, etc. to model what happens with the resonant trapping as this planet orbits - the parameters are planet mass and disk viscosity, nothing more. the planet feels the disk and the disk feels the planet, but the disk does not feel itself, ie it does not self-gravitate. Anyway, the routine isn't N-body collisional, ie the disk is like a fluid and it doesn't stick to itself. I'm a little worried about this as one would think an actual accretion disk would be an intermediate between a truly collisional model and this hydrodynamical approach. My colleagues really aren't helping much, so I thought I'd try to field the question here. Anyone have any thoughts? I know this is a rather bizarre, open ended question, but I thought I'd try.