The rand library

The math/rand library makes use of the functions found in Golang's math/rand library. For this reason, like its Golang equivalent, it should not be used for cryptographic applications.

The API is quite different from the Go version, because getting a random number from the RNG is necessarily a command.

Types

We define the following types:

Exponential = struct()
Normal = struct()
Random = struct(elements)
Seed = struct()
Shuffle = struct(elements)
Zipf = struct(s, v float, imax int)

Then the following commands are available:

Commands

put (i int) into (seed Seed)

This seeds the random number generator for when you need it to be deterministic. If you don't, it's given a strong random seed on initialization.

get (x ref) from (r Random)

This chooses a random value from the range given by the elements field of the Random value, according to the type of elements:

• If elements is a list, it selects a random element of the list. E.g: get x from Random ["one", "two", "three].
• If it's an integer i, it selects a random integer in 0::i.
• If it's int, it selects a random integer from the whole 64-bit rang of integers.
• If it's a float x, it selects a random float between 0 and x.
• If it's float, it selects a random float between 0 and 1.
• If it's bool, it selects randomly from true and false.
• If it's the name of an enum type, it selects a random element of the enum.
• If it's a pair of integers, it selects a random integer from the range.
• If it's a pair of floats, it selects a random float from the range.

get (x ref) from (s Shuffle)

This returns a random permutation of the range given by the elements field of the Shuffle value, according to the type of elements:

• If elements is a list, it returns a permutation of the list.
• If it's the name of an enum type, it returns a permutation of the elements of the enum.
• If it's a pair of integers, if returns a permutation of the integers in that range, e.g. get x from Shuffle 0::10 will get a permutation of the numbers 0 ... 9.

get (x ref) from (e Exponential)

This gets a random float greater than or equal to 0, with an exponential distribution and mean both equal to 1.

get (x ref) from (n Normal)

This returns a random float with a normal distribution having mean 0 and standard deviation 1.

get (x ref) from (z Zipf)

This gets values k ∈ [0, z.imax] such that P(k) is proportional to (z.v+k)^(-z.s). Note that in this one case, the top of the range (z.imax) is among the possibilities.