Research On Representation,Identification And Visualization Of Design Patterns In Program Comprehension

Program comprehension actually use a variety of computer application technologies, it obtains information of the program structure and design by analyzing of program source code. It can help system maintainers and developers to analyze and understand legacy systems. With the development of computer technology and the degree of expectations that users rely on computer technology to solve the problem, the size and structure of software systems is also growing. Therefore, more and more modern software engineering rely on software design patterns to solve business logic and complex software architecture. The identification of software design patterns can help software technical personnel to understand the system’s design intent and function from the software structure in the research of program comprehension and software reverse engineering.Through the research of design pattern recognition methods in program comprehension and the visual representation of design pattern recognition results, this thesis can help maintainers to understand the program. The following are the main contributions of this thesis.1. A design patterns definition language based on XML Schema, called DPDLXS, is proposed to achieve hierarchical representation of information which involves the role of design patterns of the static structure and dynamic behavior. The language we propose is the technical foundation of accurately describing design pattern features and identifying design patterns.2. An information extraction model and extraction process of source code information based on DPDLXS is proposed to extract Java source code by hierarchical structure of class static structure and dynamic behavior information. Meanwhile, a series of extraction processes and algorithms about the properties and methods in the source code are also proposed.3. Presenting the concept of correlation between the class in source code, then construct association class collection based on correlation combined with the principle of graph theory; Algorithms of design pattern identification based on correlation and feature constraints are proposed to identify the specific design pattern by feature constraint.4. According to the proposed method of design pattern identification, the method was applied to three open source applications-JUnit, JHotDraw and JRefactory. The feasibility and efficiency of this method is verified by identifying part of the design patterns and analyzing the results. Finally, complete the design and implementation of a system which is applied to identify Java language design patterns and achieve visual representation of the identification results.