CSM-424- Evolutionary Complexity: Investigations into Software Flexibility

Flexibility has been hailed as a desirable quality since the earliest days of software engineering. Classic and modern literature suggest that particular programming paradigms, architectural styles and design patterns are more “flexible” than others but stop short of suggesting objective criteria for measuring such claims. We suggest that flexibility can be measured by applying notions of measurement from computational complexity to the software evolution process. We define evolution complexity (EC) metrics, and demonstrate that— (a) EC can be used to establish informal claims on software flexibility; (b) EC can be constant ����or linear� in the size of the change; (c) EC can be used to choose the most flexible software design policy. We describe a small-scale experiment designed to test these claims.