notes_meeting_24may.txt

boulder cracking 
	- "3D" boulders
	- variables:
		+ crack year (cracks every ~10^6 years)
		+ cracking chance (mean and stdev) (as a gaussian distribution)
			-> if crack chance > 1 (shifting x_mean)
			-> add to x_mean every year (by 1/crack year)
	- when a boulder cracks, the clock "resets" on each of the resultant boulders
	- where it's going to crack:
		+ gaussian distribution 
			-> x_mean_where it cracks (0.5)
			-> sigma_where it cracks (0.1)
		+ mass breaks unevenly along diameter
		+ figuring out where to put the boulders after is tricky
			-> sometimes it falls over and sometimes it doesn't
			-> bigger piece stays where it is (with its center adjusted)
			-> smaller piece falls over (with its center adjusted, and its diameter is the "crack length" along the old diameter axis)
		+ ideally we'd store information about height?
	- we'll have to deal with the "multiple cracks" thing at some point

aeolian
	- probably ok that it's linear erosion
	- gaussian distribution
		+ mean
		+ stdev
		+ every year picks something in that distribution as the "amount eroded"
	- too small, turn into rubble pile
		+ if boulder size is < detection limit (maybe 0.25 m), remove from model

file output
	- lat_orig
	- lon_orig
	- diam_orig
	- lat_new
	- lon_new
	- diam_new
	- generation

long term
	- variables get optimized and solved for based on before/after sample data