Evaluation

Evaluation, in software architecture, is the practice of assessing whether a system or its architecture can meet its desired or required quality attributes. It comes in two complementary forms.

Qualitative evaluation is concerned with finding risks against a system’s desired qualities, rather than calculating or measuring anything directly. It works by structured judgment, surfacing risks, trade-offs, and sensitivity points in an architecture.

The Architecture Tradeoff Analysis Method (ATAM) is the best-known method for doing this systematically.

Quantitative evaluation (or quantitative analysis) measures or counts values of software artifacts, such as coupling, cyclomatic complexity, size, or test coverage. Metrics like these help identify critical or risky parts of a system, complementing the risks a qualitative evaluation surfaces with hard numbers.