What it is

Rationale is a collection of functional-programming utilities for ReasonML / OCaml, inspired by Ramda. Back in 2017 I was writing a lot of Reason, and the standard library had gaps — and worse, a lot of its list operations threw exceptions. Rationale filled in the missing pieces the way a Haskeller would want them: total functions, options instead of exceptions, and real monadic structure.

It’s the open-source project I’m proudest of. It picked up 275 stars and a handful of contributors — the most traction any of my repos has had. It’s old and unmaintained now, but people actually used it, and that still means a lot.

What’s in it

• Exception-free list operations — head, tail, init, last, nth and friends return option instead of throwing, so partial functions become total.
• Monadic Option and Belt.Result — Haskell-style functor/monad operators (<$>, >>=, …) for chaining fallible computations without a pyramid of pattern matches.
• Functional ADTs — Reader, Writer, and an IO/KIO monad. KIO carries an environment (io('a, 'env)), so you get dependency injection and isolated side effects, run once at the edge of your program.
• Point-free style — compose/pipe plus infix operators (<||, ||>), and lens composition for reaching into nested data.

Here’s the flavor — a fallible list transformation expressed as a clean monadic pipeline instead of nested matches:

open Rationale.Option.Infix;
open Rationale.Function;

RList.init(a)
  >>= (x => RList.last(a) <$> f <$> flip(RList.append, x))
  <$> RList.concat(b)
  |> Option.default(xs);

Why it still matters to me

Rationale is where my functional-programming instincts first showed up in public. The same things I cared about here — totality, isolating effects, composing small typed pieces — are exactly what pulled me toward Effect.ts and, eventually, to building Stax years later. Different language, same values.

Links

• Source: github.com/jonlaing/rationale