Refactoring Principles, Code Smells, and Refactoring Techniques

Refactoring is the process of improving the internal structure of software without changing its observable behavior, aiming to increase understandability and reduce modification cost.

Key motivations include improving design, uncovering bugs, and speeding up development; refactoring should be done continuously rather than as a separate task, following the "three‑rule" guideline.

The article lists common code smells such as duplicated code, long or large classes, long parameter lists, divergent changes, shotgun surgery, data clumps, primitive obsession, switch statements, parallel inheritance hierarchies, redundant classes, speculative future features, confusing temporary fields, excessive coupling, middle‑man, and more.

For each smell, concrete refactoring techniques are described, e.g., extract method, inline method, extract temporary variable, replace algorithm, move function, move field, extract class, inline class, hide delegation, remove middle‑man, introduce extension functions, etc., with illustrative Java snippets.

Additional sections cover reorganizing data (encapsulating fields, replacing data values with objects, converting value objects to references, using parameter objects, removing setters), simplifying conditional expressions (splitting, merging, removing control flags, replacing with polymorphism), simplifying function calls (renaming, adding/removing parameters, separating queries from commands, introducing parameter objects, hiding functions), and handling generalization relationships (pull‑up fields/methods, push‑down, extract super‑class/interface, replace inheritance with delegation, etc.).