Mining Software Quality at Design Phase by Recognizing Architectural Patterns and Computing Pattern Software Metrics

Abstract

Clients need Software to be developed within few days and that too of good quality. And here lies the need for reusability, not only at code level but also at model level. This reusability can be achieved with the help of the use of Design Patterns, which further speeds up the process of product development by providing the pre-tested footsteps to follow. Design patterns show their effectiveness from design phase to maintenance. We propose that the quality of the software design can be determined by presence of Design patterns, which symbolizes good arrangement of data and control as a part of software structure. The software structure contributing to a design pattern should form a cohesive unit serving towards a single purpose and should pose less amount of coupling with the rest of the software elements. In this paper, we provide a metric based study of assessing quality of software structure by detecting the presence of the design patterns and finding the amount of cohesion and coupling of each design pattern. Higher is the quality of the used design pattern higher will be the quality of the product, with more flexibility towards change and maintenance. An early feedback at the design phase regarding software design quality would help in risk analysis as well as quality assurance.