Testability

Testability is the degree of effectiveness and efficiency with which test criteria can be established for a system, product, or component, and with which tests can then be performed to determine whether those criteria have been met. It is a quality attribute and, per ISO/IEC 25010, a sub-characteristic of maintainability. A system that’s hard to test is, by the same token, hard to change with confidence.

A system’s testability is shaped largely by its design rather than by its tests. High cohesion and low coupling let a component be exercised in isolation, without dragging in the rest of the system as fixtures or mocks. Dependency injection and programming to an interface make it possible to substitute real collaborators with test doubles, and observable, deterministic behavior — pure functions, explicit inputs and outputs, no hidden global state — makes test criteria easier to state in the first place.

Testability is what makes disciplines such as test-driven development and unit testing practical at scale. Where a codebase resists testing, the usual remedy is refactoring toward smaller, more loosely coupled units rather than writing ever more elaborate test setup around the existing structure.

References

• ISO/IEC (2023). ISO/IEC 25010:2023 — Systems and software Quality Requirements and Evaluation (SQuaRE) — Quality model.