pure-functions.md

Benefits of pure functions

1. Easier to reason about

   • a pure function tells everything you need to know about the function. No hidden I/O, no external variables modified, no instance or even global variables to hunt down.

2. Easier to test

   • since the same input(s) always return the same output(s), they're easy to test. Provide the input and expected output. Simple! No need for setup, teardown, or mocking.

   expect(last([3, 9, 5])).toBe(5) // that's it!

3. Easier to debug

   • Debugging is like detective work, and there are less suspects to investigate, since you don't have to worry about mutations, side-effects, and variables outside the pure function.

4. Easier to combine

   • Pure functions can act like LEGO blocks, when used with function composition or piping.

5. Easier to parallelize

   • Since pure functions don't modify any of its parameters or any outside variables, you can be sure they're thread-safe. The OS can run these functions using any number of cores and in any order.

6. Easier to cache

   • Some functions are expensive to calculate, e.g. the permutations of abcde. The results of a pure function can be cached, and calling permutation('abcde') the second time will be faster, since the result will be retrieved from the 🐇cache rather than 🐢recomputed.

7. and many other benefits

   • Referential transparency, laziness, idempotency,... It even makes it easy for bundlers to tree-shake frontend code that's composed of pure functions.