"IT Isn't Different After All": Implications of Normalized Systems for the Industrialization of Software Development

Evolvability has emerged as a crucial non-functional requirement for software. Normalized Systems (NS) theory studies how modular structures in software respond to increasing amounts of change. The theory is based on concepts such as stability from systems theory and consists of the derivation of four principles, which locate and identify so-called combinatorial effects. The theory also shows how code generators (called expanders) can be built of which the absence of combinatorial effects towards anticipated changes can be proven. The implications of NS theory for the large-scale production of software are discussed. First, NS theory shows that addressing the high demands in terms of evolvability of software, requires new levels of fine-grained modularity in theory and practice. Second, NS theory shows that this modularity can only be achieved by highly structured and automated software development processes. Several steps towards realizing this in practice are discussed. It is concluded that by applying concepts from traditional engineering such as in NS theory, new levels of fine-grained modularity can be reached to move closer to industrialization of software development, thereby answering the ever-increasing call for more evolvability in software.