## Randomness in the style of Common Lisp

This proposal for random numbers is not a literal transcription of the Common Lisp (CL) interface, but it uses the same concepts and provides the same facilities. The text is a heavily edited version of the CL Hyperspec.

A *random-source object* is an encapsulation of the state information used by an implementation-dependent pseudo-random number generator. The state can be printed out and successfully read back in by the same implementation, but might not function correctly as a state in another implementation. Random-source objects are a disjoint type. This facility should not be used where strong randomness is required.

## Procedures

`(make-random-source . `*args*`)`

Constructs and returns a random-source object that has been randomly initialized by some implementation-defined means. The CL equivalent is `(make-random-state t)`. If *args* are provided, they are used in an implementation-defined way to affect what result is returned.

`(random-source? `*obj*`)`

Returns `#t` if *obj* is a random-source object, and `#f` otherwise. The CL equivalent is `random-state-p`.

`(random-source-state `*random-source*`)`

Returns an implementation-specific object representing a copy of the state encapsulated by *random-source*. This object MUST be printable and rereadable using standard Scheme lexical syntax. It MUST also be suitable for passing to `make-random-source-from-state`. Providing this mechanism makes it possible to save and reconstitute a random-source in a file or database, or to pass it across a network to an equivalent implementation. Mutating the result of this procedure does not affect *random-source*. There is no CL equivalent of this procedure, because CL `random-state` objects are themselves required to be printable and rereadable (they are typically CL `structs`).

`(make-random-source-from-state `*state*`)`

Constructs and returns a random-source object whose state is a copy of *state*, so that mutating *state* does not affect the random-source object. The result will generate the same sequence of pseudo-random numbers that the original random-source object would have generated as of the time `random-source-state` was invoked on it. It is an error to pass a *state* object that has been mutated. There is no CL equivalent of this procedure.

`(copy-random-source `*random-source*`)`

Constructs and returns a random-source object whose state is an independent copy of the state of *random-source*. Calling this procedure is equivalent to calling `(make-random-source-from-state (random-source-state `*random-source*`))`, but potentially more efficient because it can avoid copying the state twice. The result and *random-source* will henceforth return the same sequence of values, allowing the same series of pseudo-random numbers to be generated many times within a single program. The CL equivalent is `make-random-state` with a random-state argument.

`(current-random-source`)`

A parameter that returns the default random-source object used by `random`. Its initial value MUST be a random-source object, but is implementation-dependent. The CL equivalent is `*random-state*`.

Note: The equivalent of CL `(make-random-state)` and `(make-random-state nil)` is `(copy-random-source (current-random-source))`.

`(random `*limit* [*random-source*]`)`

Returns the next pseudo-random number from *random-source*. The result is a non-negative number less than *limit*. *Limit* MUST be either an exact integer (in which case `random` returns an exact integer), or an inexact real number (in which case `random` returns an inexact real number). If *random-source* is not specified, the value of `(current-random-source)` is used. The CL equivalent of this procedure is `random`.

An approximately uniform choice distribution is used. If *limit* is an integer, each of the possible results occurs with (approximate) probability 1/*limit*.

## Examples

(<= 0 (random 1000) 1000) => true (let ((state1 (copy-random-source (current-random-source))) (state2 (copy-random-source (current-random-source)))) (= (random 1000 state1) (random 1000 state2))) => true