Using system dynamics to teach about dependencies, correlation and systemic thinking on the software process workflows

Abstract

It is important to count on tools to help software professionals to evaluate the software process and how it may be affected by factors related to its deployment. Simulation models are a valuable means to illustrate the behaviour of such a process since scenario generation supports the prediction of potential outcomes and the prevention of undesired scenarios which are harmful to the process and the company in charge of the project to be developed. This work explores the effectiveness of introducing system dynamics (SD) models in the software engineers’ process of understanding, from a management perspective, the software process dynamics. The used SD simulation model of the software process emphasises the representation of an iterative process. The COCOMO II model drivers and their main attributes were used, providing a set of reference factors that affect the software process, the estimation of project cost and the effort required. A set of 59 junior software professionals with no previous knowledge about SD participated in a validation study. For simple predictive scenarios, there was no important improvement effect, while for more complex predictive scenarios SD helped them to guess better and provide a rationale for the expected behaviour of the software process performance.