Code Smell

Also Known As

TL;DR

Explanation

The term was popularised by Martin Fowler in "Refactoring". A code smell is not necessarily broken — the code may work — but it suggests structural issues that will make the code harder to change safely over time. Common smells: long methods, large classes, long parameter lists, feature envy, data clumps, primitive obsession, and duplicate code. Smells are signals to investigate, not mandates to refactor immediately.

Common Misconception

Why It Matters

Common Mistakes

• Treating all code smells as equally urgent — prioritise smells in frequently-changed, business-critical code.
• Refactoring smells without tests in place — you need a safety net before restructuring.
• Conflating code style violations (formatting) with structural smells — they require different remediation.
• Using smell detection tools as a pass/fail gate rather than a signal to investigate.

Code Examples

// Multiple smells in one function:
function p(\$d) {  // cryptic name
    global \$db;   // global state
    if(\$d['t']==1) { // magic number
        \$r = \$db->query("SELECT * FROM orders WHERE uid=".\$d['u']); // SQL injection
        foreach(\$r as \$row) { echo \$row['total'] * 0.9; } // magic number
    }
}

function applyLoyaltyDiscount(array \$request): void {
    \$orders   = \$this->orderRepo->findByUserId(\$request['user_id']);
    \$discount = LoyaltyDiscount::RATE; // named constant
    foreach (\$orders as \$order) {
        echo \$order->total * (1 - \$discount);
    }
}

References

• ↗ https://refactoring.guru/refactoring/smells
• ↗ https://martinfowler.com/books/refactoring.html

Tags

Related Terms