<importantQuestionIsItPracticalToUseArraysForAllNodes>
	
	<question>
		Is it practical to use arrays for all nodes?
		<context>
			The design I'm planning is for all nodes in a network,
			and all lists of nodes, to be Object arrays,
			and for all floating points in the network to be in double arrays
			in those Object arrays.
			The whole shape and size of the network is represented that way.
			A network can have other types of things, but network is more complex than node.
		</context>
	</question>
	
	There are 2 main ways of representing nodes and node lists:
	(1) As arrays, or (2) as variable size lists.
	
	Floating points can be modified without changing size.
	
	Swapping the position of 2 childs can theoretically be done both ways.
	
	ARRAY: Adding 1 child to a neural node with 10000 childs has a cost of 10001
	LIST: cost of 1
	
	But that is not as relevant because merging nodes can add many childs at once for the same cost.
	
	The main reason ARRAY is fast enough, despite its slowness of changing network shape,
	is that the network shape changes shape much less often than it executes.
	
	If the network goes through an average of 100 iterations of half its node
	every time half its nodes gain or lose child node(s), then its probably fast enough.
	
</importantQuestionIsItPracticalToUseArraysForAllNodes>