Specification and use of concern metrics for supporting modularity-oriented modernizations

Abstract

Bad modularized concerns are a known deficiency of legacy systems, making their maintenance increasingly harder and expensive. An alternative is to conduct a modularity-oriented modernization (MOM), whose goal is to restructure concerns for making them more modular. Architecture-driven modernization (ADM) is a model-based alternative for that—it is a kind of reengineering that employs standard metamodels along the process. Knowledge Discovery Metamodel (KDM) is the main one, and it is used for representing all aspects of an existing system to be modernized. ADM also delivers a metamodel called Structured Metrics Metamodel (SMM) whose goal is to standardize the specification of metrics, measurement processes, and results of the measurements. A central task in any modernization process is to measure, characterizing quantitatively, the legacy and the modernized version to check if problems were solved/assuaged. In a MOM, it is important to measure to what extent the concerns are spread across the system using concern metrics. However, there are no studies that investigate how SMM can be used to specify concern metrics in MOMs. Therefore, we present our approach of how to use SMM for specifying concern metrics in a modularity-oriented modernization in the ADM context, i.e., using KDM and SMM. We detail the specification of two concern metrics: Concern Diffusion over Operations (CDO) and Concern Diffusion over Components (CDC). Besides, we also present a tool that can be used in a MOM to support the measurement process