Towards a model to support in silico studies of software evolution

Software evolution is recognized as one of the most challenging areas in the field of Software Engineering. The observation of evolution is time-dependent, reducing opportunities for actual observations in short periods of time. Usually, maintenance cycles are proportional to the software life cycle. Therefore, the amount of research has not been enough to deal with all the issues related to the evolution of software. However, simulation through confident models represents an interesting strategy to support software decay observation in short period of time. Towards that, this paper describes a model aimed at supporting the software decay simulation through systems dynamics. The Laws of Software Evolution and ISO 9126 were used as initial knowledge to support the discovery of software characteristic (size, periodicity, complexity, effort, reliability, and maintainability) relationships. Next, evidence to strengthen the existence of such relationships was acquired through quasi-systematic literature reviews. In sequence, the model was applied to support the simulation of industrial software decay. The results suggested its feasibility and correctness, making it an interesting candidate to support future software decay studies.