Favor Pure Functions and Explicit Dependencies for Maintainable Code

One of the most impactful habits for maintainability is writing pure functions whenever possible — functions that depend only on their inputs and produce no side effects. Pure functions are trivial to test, easy to reason about, and safe to refactor or parallelize. When you do need side effects (database calls, API requests, file I/O), push them to the edges of your system and inject dependencies explicitly rather than reaching for global state or singletons. This makes data flow obvious and reduces the cognitive load when debugging or modifying code months later.

A practical way to enforce this is to structure modules around a clear "core" of pure logic and a thin "shell" that handles I/O and wiring. For example, instead of embedding a database call inside a business logic function, pass the required data as arguments or accept a repository interface. This lets you unit test the core logic with simple fixtures and swap implementations (e.g., a mock for tests, a real client in production) without touching the logic itself. The result is code that’s easier to read, test, and change — exactly what maintainability demands.