Third (and likely last) iteration on the "galaxy collision toy model". I randomized a bit the initial conditions and increased the number of test masses to avoid visual artefacts.
If one day I decide to learn how to implement an octree I might come back to this 
pic.twitter.com/JDhKlzd73Z
-
Circular orbit + small perturbation as initial conditions for the test masses. Looked like the easiest approach.
