PlasmaPong - Realtime CFD, comments?
 

If you haven't played with it already you're in for a pleasant surprise:

http://www.plasmapong.com/

Stimulated by this work:

http://www.dgp.toronto.edu/people/st.../pdf/GDC03.pdf

It certainly appears to produce what it sets out to, i.e. a visually realistic realtime fluid flow prediction

The approach appears to be a hybrid Eulerian/Lagrangian one. I haven't delved deep enough yet to see where the assumptions (and there must be big ones!!) come into play. Has anyone else?