lowlife

An implementation of Conway's Game of Life. Work-in-progress highly unstable interfaces may change without warning use at your own risk here there be dragons.

Usage

lowlife's data model is based on (immutable) sets of Cartesian coordinate pairs, which represent the locations of all living cells on the grid.

import * as Life from 'lowlife'

//  . O .
//  . . O   glider (origin at top left)
//  O O O

let glider = Life.FromLiving([[1,0], [2,1], [0,2], [1,2], [2,2]])
let iteratedGlider = glider
for (let i = 0; i < 4; i++)
  iteratedGlider = Life.Next(iteratedGlider)

assert(
  JSON.stringify([...Life.Living(iteratedGlider)])
  ===
  JSON.stringify([...Life.Living(glider)].map(([x, y]) => [x + 1, y + 1]))
)

The Next function can optionally reuse a grid's memory when constructing its successor. Any operations on the original grid after using this option will throw an exception:

let glider = Life.FromLiving([[1,0], [2,1], [0,2], [1,2], [2,2]])
let iteratedGlider = glider
for (let i = 0; i < 4; i++)
  iteratedGlider = Life.Next(iteratedGlider, {canFree: true}) // <- Tells `Next` to reuse its argument's memory

assert(
  JSON.stringify([...Life.Living(iteratedGlider)])
  ===
  JSON.stringify([...Life.Living(glider)].map(([x, y]) => [x + 1, y + 1]))
)

This time Living(glider) throws a TypeError. Since iteratedGlider was initially an alias of glider, glider was revoked after the first call to Next

Architecture

Modules, in ascending order of abstraction:

• branch and leaf provide quadtree operations for their respective subtypes
• tree assembles them (while applying canonicalization/memoization transforms) to provide full quadtree operations
• grid embeds a quadtree in the coordinate plane, growing when needed to simulate infinite size
• life provides memory management