Recipe

Build a function test-first

Use when: you are about to write a function whose behaviour you can describe with examples

Goal: a working function where every behaviour has a test that was seen to fail first

Taught in Fixtures, edge cases, and TDD

1. 1

   Choose the simplest thing the function should do that it doesn't do yet.

   Example

   "900 cents among 3 people is three shares of 300."

   Why: small steps keep each failure easy to understand and each fix small.

2. 2

   Write one test for that behaviour and run it; watch it fail for the reason you expect.

   Example

   assert split_bill(900, 3) == [300, 300, 300] fails with assert None == [...].

   Why: a test you've seen fail is a test that can fail.

3. 3

   Write the simplest code that passes this test, then run every test.

   Example

   return [total // people] * people.

   Why: the simplest version shows exactly what the tests demand, and no more.

4. 4

   Improve names and structure without changing behaviour; re-run the tests after each change.

   Example

   Add the docstring; the two tests stay green.

   Why: the tests prove the clean-up changed nothing they check.

5. 5

   Go back to step 1 with the next behaviour, until the edge cases are covered.

   Example

   split_bill(100, 3) == [34, 33, 33], then split_bill(10, 0) raising ValueError.

   Why: each round adds one tested behaviour, so the suite grows with the code.