A Method for Merging Experts' Cause-Effect Knowledge in Software Dependability

Software dependability engineering is an interdisciplinary area built on numerous concepts with complex interrelations. Understanding the relations between various concepts and integrating such conceptual knowledge is fundamental for educational, research and industrial purposes. This paper proposes a method to merge experts' conceptual knowledge represented in the form of Causal Mechanism Graphs (CMG). A case study was conducted to apply the method on 14 causal mechanism graphs obtained from 11 domain experts. The obtained consensus knowledge was then validated and evolved based on another 24 experts' opinions. The results demonstrate the main causal mechanisms that influence software dependability attributes, i.e. software aspect of reliability, safety, security, availability and maintainability. The application shows that the CMG merging method has the advantage of explicitly aggregating complex causal knowledge without losing information on the original knowledge structure. The application also shows that the CMG merging method is capable to integrate various factors that influence multiple dependability attributes at different stages of the software development lifecycle.