L2: L2-error
In ezer/NITPicker: Finds the Best Subset of Points to Sample

Given two functions y1(t) and y2(t), this function finds the L2-distance between the following two curves: a) y1(t)-y2(t) sampled at all time points (tp) b) y1(t)-y2(t) sampled at the time points indexed by index (tp[index]). Note that by setting y2 to rep(0,length(tp)), this function can be used to estimate the L2-error in the shape of y1.

1	L2(tp, y1, y2, start, stop, index, numSubdivisions = 2000)

`tp`	A numerical vector of time points (or spatial coordinates along a single axis)
`y1`	A numerical vector of measurements (of the control)
`y2`	A numerical vector of measurements (of the experimental condition)
`start`	A numerical value representing the start time (or spatial coordinate) of the integration
`stop`	A numerical value representing the end time (or spatial coordinate) of the integration
`index`	A vector of positive integers representing the indices of `tp` that we subsample
`numSubdivisions`	This can be adjusted to ensure the integration doesn't take too long, especially if we aren't overly concerned with rounding errors.