Skip to content


Subversion checkout URL

You can clone with
Download ZIP
A suite of Haskell libraries for representing, manipulating, and sampling random variables
Fetching latest commit...
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
rvar Bump version.
tests Make benchmarks build
.gitignore Added a .gitignore Add some basic installation instructions in README


Random-number generation based on modeling random variables as first-class entities. This library currently consists of 3 major parts:


This package defines the central datatype of the system, a monad transformer called RVarT which extends an arbitrary monad with non-backtracking nondeterminism. In particular, the RVar type (an alias for RVar Identity) models pure random variables.


This package provides the backend for RVarT; arbitrary sources of entropy. Its design is still in major flux.


This package provides an end-user interface that defines random variables following several standard distributions as well as some convenient interfaces for sampling them.


To use the system, you'll typically want import at least two modules: Data.Random for the main interface and a supported entropy source, such as System.Random.MWC from the mwc-random package. You may also want to import one or more of the extra distributions provided in the Data.Random.Distribution heirarchy (uniform and normal are exported by Data.Random automatically). Then, you can define random variables using do notation, sample them using sampleFrom, etc. For example:

import Data.Random
import System.Random.MWC (create)

logNormal :: Double -> Double -> RVar Double
logNormal mu sigmaSq = do
    x <- normal mu sigmaSq
    return (exp x)

main = do
    mwc <- create
    y <- sampleFrom mwc (logNormal 5 1)
    print y


Get the latest release from Hackage:

cabal install random-fu

Or a bleeding-edge version from github:

git clone
cd random-fu
(cd random-source; cabal install)
(cd rvar;          cabal install)
(cd random-fu;     cabal install)
Something went wrong with that request. Please try again.