[lpvwwuza] Sampling random distributions

One of the fundamental things computers can do extremely well compared to humans is generate randomness, technically pseudorandomness.

Uniform real between 0 and 1.
Uniform discrete.
Weighted discrete (categorical).
Fixed number of draws from a weighted discrete distribution (possibly with duplicates of varying weight) without replacement, with or without ordering.
Permutation.
Even permutation (alternating group).
Circular permutation.
Subset (power set).
Bipartite matching.
Normal.
Exponential.
Gamma (?).
Multivariate Gaussian with covariance matrix.
Point inside a hypersphere (rejection sampling from a hypercube becomes very inefficient in high dimensions).
Point on the surface of a hypersphere.
Hyperellipse (hyperellipsoid) instead of hypersphere (analogous to Gaussian with covariance).
Element of a orthogonal group.
Special orthogonal group (hyper rotation).
Unitary group (?).

Previous fundamental computer ability (to pay attention).

Inspired by the convenience of the rexp function in R.