The test-driven development loop

Goal

Grow code in small, verified increments so that every behaviour is specified by a test before it is implemented and the design is refactored continuously while tests protect it.

Prerequisites

A fast test runner (seconds, not minutes) and a unit small enough to test in isolation.

Steps

1. Write one test for the next small behaviour and run it; it must fail for the expected reason.
2. Write the minimal code that makes the test pass, even if it looks naive.
3. Run all tests; with everything green, refactor duplication and unclear names without changing behaviour.
4. Repeat. Keep the loop short; if a step takes more than a few minutes, the increment is too large.
5. Commit at green points with messages that state the behaviour added.

Expected result

A suite that documents behaviour step by step, code that has no untested branches, and a design that was refactored while it was still cheap.

Limits and test basis

The loop suits logic with clear inputs and outputs; user-interface layout or exploratory prototyping fit it less well. TDD does not guarantee good design by itself; the refactoring step is where design happens and it is the step most often skipped. The description follows the cited article.