Tdd

TDD Skill

What You Get

• Failing test (RED) -> minimal implementation (GREEN) -> refactored code, committed at each green step
• Full test suite passing with no mocks of internal modules

Philosophy

Tests verify behavior through public interfaces, not implementation details. Never test private methods directly. Never mock your own modules.

Use vertical slices, not horizontal slices. Write one test, make it pass, repeat. Do not write all tests first then implement -- that is waterfall disguised as TDD.

Workflow: 4 Phases

Phase 1: Planning

Before writing any code, confirm:

• What is the public interface (function signatures, module API)?
• What behaviors does the caller expect?
• What are the edge cases and error conditions?

Do NOT write code yet. Agree on the interface first.

Phase 2: Tracer Bullet

Write exactly ONE test for the simplest happy-path behavior. Run it. Watch it fail (RED). Implement the minimum code to make it pass (GREEN). This proves the test infrastructure works end-to-end.

Phase 3: Incremental Loop

For each remaining behavior:

1. Write ONE failing test (RED)
2. Implement minimum code to pass (GREEN)
3. If the code is messy, refactor NOW while green
4. Run the full suite -- all tests must pass
5. Repeat

Never skip ahead. Never write two tests before making the first green.

Phase 4: Refactor

Only refactor when GREEN. Look for: duplication, shallow modules, long methods, feature envy, primitive obsession. Run the full suite after every refactor step.

WRONG: writing all tests first

# WRONG: this is waterfall disguised as TDD
def test_add(): ...
def test_subtract(): ...
def test_multiply(): ...
def test_divide(): ...
def test_divide_by_zero(): ...
# then implement all of Calculator at once

CORRECT: one test at a time

# RED: write one failing test
def test_add():
    assert Calculator().add(2, 3) == 5

# GREEN: implement minimum to pass
class Calculator:
    def add(self, a, b): return a + b

# next test only after green
def test_subtract():
    assert Calculator().subtract(5, 3) == 2

Cycle Checklist

• Test describes a behavior, not an implementation detail
• Test uses the public API, not internal methods
• Test is RED before writing implementation
• Implementation is the minimum to go GREEN
• No new mocks of internal modules (only system boundaries)
• Full suite passes after going GREEN
• Refactor only while GREEN, re-run suite after each change

Sub-files

See also

• property-testing -- property-based testing (Phase 3.5 after TDD loop)
• refactoring-strategy -- large-scale refactoring beyond post-TDD cleanup