reduce-map-filter.js

const reverse = (array) =>
  array.reduce((acc, elem) => [elem].concat(acc), []);

const map = (array, fun) =>
  reverse(array.reduce((acc, elem) => [fun(elem)].concat(acc), []));

const filter = (array, pred) =>
  reverse(array.reduce((acc, elem) => pred(elem) ? [elem].concat(acc) : acc, []));

The array reversing in map and filter could be skipped by concatenating with the spread syntax.

const map = (array, fun) =>
  array.reduce((acc, elem) => [...acc, fun(elem)], []);
  
const filter = (array, pred) =>
  array.reduce((acc, elem) => pred(elem) ? [...acc, elem] : acc, []);

@forsureitsme I mean, yeah, sure, but I wanted to simulate a situation where I'd have nothing but cons. The reason I'm using JS here is for accessibility. Concatenation is a non-trivial op if you're working with lists.

These implementations aren't meant to be good or efficient, just show that even if all you can do is prepend you can build map and filter just using reduce.

Also, these are theoretical implementations! We can do a lot better than either yours or mine if we actually want to build something JS-friendly (probably using internal mutation) :)