Check calls to children() aren't assuming too much
At least the MC_TAUS and MC_TAU_Decay routines were using children() where stableDescendants() was really the correct function. I'm opening this issue to make sure we consider other analyses using children() are doing The Right Thing.