Abstract The Pattern of the Viable System and its Language

The Viable System Model tries to capture the essential pattern of living systems: both natural organisms and human organizations. A Viable system is one that can survive in its environment. Survival in a dynamic environment requires adaptation. The Viable System Model specifies a supervisory-adaptive-control or “intelligent ” control strategy as necessary and sufficient for survival. This paper concentrates on the technical aspects of software viability. Software viability is defined as the quality a software system has if its architecture can be adapted over time by humans (adaptable at design-time) toward becoming an “intelligent ” control system (adaptive at run-time). The Viable System Model is described using Alexander’s pattern language form and related to software architecture. This pattern language shows how software may be evolved to be viable as defined above. The language is presented as a sequence of patterns from large, embracing to smaller, embraced patterns. Seven patterns describe the sequence of the viable system. Two further patterns are required to make it a generative language. All the patterns are present in current, viable, software systems. However, humans are performing the higher order functions of control. Thus, the evolution of a software system in this approach is characterized by the migration of human control skills into software until the system reaches the desired state of autonomy