Curious what people think about this implementation of Conway's Game of Life in Haskell.

{-# LANGUAGE QuasiQuotes #-}

module Test.Conway (
  ) where

import           Data.Array.Repa
import           Data.Array.Repa.Stencil
import           Data.Array.Repa.Stencil.Dim2
import           Data.Functor.Identity
import           System.Random

conwayNeighborsStencil :: Stencil DIM2 Int
conwayNeighborsStencil =
  [stencil2| 1 1 1
             1 0 1
             1 1 1 |]

conwayStep :: Array U DIM2 Int -> Array U DIM2 Int
conwayStep r =
  runIdentity . computeP . szipWith rules r .            -- Apply rules
  mapStencil2 (BoundConst 0) conwayNeighborsStencil $ r -- Find # neigbors
    rules 1 2 = 1 -- Alive and has two neighbours
    rules _ 3 = 1 -- Alive or dead and has three neighbours
    rules _ _ = 0 -- All others are dead

conwayMakeRandom :: RandomGen g => g -> DIM2 -> Array U DIM2 Int
conwayMakeRandom g sh =
  fromListUnboxed sh $ take (size sh) $ randomRs (0, 1) g

Naturally, conwayMakeRandom creates the initial grid and conwayStep advances the evolution a grid.


That's a nice use of stencil2 and Repa. All in all there isn't much to review, it follows a common approach on Conway's Game of Life with Repa. No type signatures are missing, but with Repa it's more or less mandatory to add them.

Well done. But the next time, consider adding all your code, conwayCalculateAge and conwayMakeEmpty were probably interesting (I suspect conwayMakeRandomIO to be the same as conwayMakeRandom but using randomRIOs = randomRs (0, 1) <$> newStdGen).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.