LOOP, FEEDBACK, and RAND nodes

[jackparmer]

LOOP, FEEDBACK, and RAND nodes

1. The LOOP example in the docs pages is too complex on its own. There needs to be another example before it that is as simple as possible -- not more than 3 nodes. There needs to be a reasonable, well-written explanation of what is happening in both apps.
2. The LOOP node should be moved from Simulations to Logic/Flow
3. The Feedback node needs a docstring
4. The FEEDBACK node needs another example that is as simple as possible - no more than 3 nodes. There needs to be a reasonable, well-written explanation of what is happening in both example apps.
5. The level of zoom is off in the existing LOOP and FEEDBACK examples - it's so zoomed into the app that the app is cut off
6. The RAND node should be modified so that it doesn't necessarily need an input. If no input is provided, it should just return a data container of type scalar (previously called "constant") that contains a single random number generated by RAND. there should also be a parameter than lets you change the distribution (uniform, gaussian, poisson, dirichlet)
7. Once 1-6 are done, I'd like to see a super simple demo app that a) generates a random number once per second, 2) adds it to a histogram, 3) repeats 100-1000 times. if RAND is set to gaussian, the histogram should obviously converge towards a bell curve