I wish the `loc` and `scale` arguments of `rvs_normal` were optional when generating a 1D array of variates, so that one could write `x = rvs_normal(array_size=n)` instead of `x = rvs_normal(0.0_dp, 1.0_dp, array_size=n)` That is how numpy.random.normal works.