I was planning on, as a computer science project, simulating the collapse of
a turbulent gass cloud and the formation of starts, but I have come to
beleive the model is too complex and astronomy bound for me right now.

I would like to hear some things on the required complexity of a SPH model
of the creation of the moon from an impack into proto earth which threw out
debris into orbit. From my, non expert, point of view it seems that this can
be tested quite easily, but... what are the features that the model should
exhibit when testing it?

Am I being overly optimistic when thinking about this problem as two
objects, consisting of a number of particles of different type (different
materials) colliding and then this is just a n-body problem which does or
does not over time lump back into two new objects, one orbiting the other?
Which parameters should be included and how do I validate the results
afterwards?