Based on the paper at http://arxiv.org/pdf/gr-qc/9609015v2.pdf

This new sim differs from my old one in that the old one used finite differencing. This new one is based on hyperbolic formalisms. My hydrodynamic solvers here and here had been a warm-up.

This one is written in LuaJIT, the best prototyping/performance scripting language in existence. I've incorporated my Symbolic Lua code such that the initial conditions can be provided as equations and the system performs all the differentiation necessary. Some day the Roe matrices will be deduced from simply providing the PDEs.