Formal Verification, Quantitative Analysis and Automated Detection of Design Patterns

Abstract

Present-day software engineering concepts emphasize on developing software based on design patterns. Design patterns form the basis of generic solution to a recurring design problem. Software requirement analysis and design methodologies based on different Unified Modelling Language (UML) diagrams need to be strengthened by the use of a number of design patterns. In this study, an attempt has been made for automated verification of the design patterns. A grammar has been developed for verification and recognition of selected design patterns. ANTLR (ANother Tool for Language Recognition) tool has been used for verification of developed grammar. After proper verification and validation of design patterns, there comes a need to quantitatively determine the quality of design patterns. Hence, we have provided a methodology to compare the quality attributes of a system having design pattern solution with a system having non-pattern solution, both the system intending to provide same functionalities. Using Quality Model for Object-Oriented Design (QMOOD) approach, the cut-off points are calculated in order to provide the exact size of the system in terms of the number of classes, for which the solution adopted using design pattern, provides more quality parameters. Again Design Pattern Detection (DPD) has also considered as an emerging field of Software Reverse Engineering. An attempt has been made to present a noble approach for design pattern detection with the help of Graph Isomorphism and Normalized Cross Correlation (NCC) techniques. Eclipse Plugin i.e., ObjectAid is used to extract UML class diagrams as well as the eXtensible Markup Language (XML) files from the Software System and Design Pattern. An algorithm is proposed to extract relevant information from the XML files, and Graph Isomorphism technique is used to find the pattern subgraph. Use of NCC provides the percentage existence of the pattern in the system.