Engineering Craft › Clean Code & Principles

Dead Code

Code that never runs and should be deleted.

Also known as: unused code, unreachable code

Dead code is code that never runs or whose result is never used: unreachable branches, unused functions, variables nobody reads, old feature-flag paths, commented-out blocks.

def total(items):
    legacy_total = sum(i.old_price for i in items)    # never used
    return sum(i.price for i in items)

    print("done")                                      # unreachable, after return

# def old_discount(order): ...                         # commented out "just in case"

Why remove it

• It costs reading time. Every reader wonders whether it matters.
• It misleads, and rots. It’s no longer tested or maintained, so it quietly becomes wrong.
• It hides real code in search results and complicates refactoring.
• It carries risk: dead code can still hold old vulnerabilities, or be revived by accident.

Why it’s safe to delete

Version control remembers everything. If you ever need it, git log and git show bring it back, so don’t comment out code to keep it (Git).

Finding it

• Linters and editors flag unused variables, imports and unreachable statements (linters).
• Coverage reports show lines that no test reaches (test coverage), though untested isn’t always dead.
• Search for usages before deleting a function, including dynamic calls, reflection, configuration and other repositories.
• Logs or metrics can show whether an endpoint is still hit in production.
• Old feature flags leave dead branches. Remove them when a rollout is done (flag cleanup).

Deleting carefully

• Remove it in a small, separate commit, so it’s easy to revert.
• Run the tests, and check that nothing refers to it dynamically.
• For public APIs, deprecate first, because outside users may depend on it (deprecation).

Deleting code is a legitimate, valuable contribution. Less code means fewer bugs (boy scout rule).