# Haskell Monad Chaining

This note covers `35-haskell-monad-chaining/`, which sequences several dependent rollout checks with `Either` and `do` notation.

---

## 1. Why Monad Chaining Matters

Some workflows cannot be expressed as independent field checks.

In this example, later steps depend on earlier successful results:

- find the service profile first,
- then check whether the environment is allowed,
- then validate the production change ticket,
- then choose the image tag for the requested track, and
- finally choose the approver.

That dependency chain is the point. Each step needs the result of the previous one.

---

## 2. Why `Either` Still Fits

This example does not need accumulated errors. It needs short-circuiting business logic.

That makes `Either String` a good fit:

```haskell
approveRollout :: Catalog -> RolloutRequest -> Either String ApprovedRollout
```

`do` notation keeps the happy path readable while preserving the fail-fast semantics.

---

## 3. How This Complements the Validation Example

`28-haskell-applicative-validation/` teaches independent checks that all run so several errors can be reported together.

This example teaches the opposite shape:

- one decision unlocks the next, and
- the workflow stops once a prerequisite fails.

That contrast is useful. It shows why “Applicative versus Monad” is not just theory. The control flow shape changes the design.

---

## 4. Commands to Try

```bash
cd 35-haskell-monad-chaining

nix develop
cabal run
cabal run -- api:production:stable:CHG-2048
cabal test

nix build
./result/bin/mini-monad-chain api:production:stable:CHG-2048

nix run . -- api:production:stable:CHG-2048
nix flake check
```