Research On Information Extraction And Visualization In Program Comprehension

Program comprehension is a process that abstracts the high-level program descriptionactually. It gets the pattern of program comprehension, including the real world objects and therelationship among them, and ignoring some unimportant information of the program design.Program comprehension begins with the existing code and uses computer technology to realizethe code. It analyzes the program structure automatically and gets related knowledge of program,then reconstructs the software structure, algorithm description and so on. Programcomprehension can help technical personnel to understand the software effectively.This paper bases on the theory of program comprehension, and proposes an algorithm aboutcomprehension and description object-oriented program. This methodology expresses theprogram structure in visual form. The first step is to analyze object-oriented program source code,with the code analysis of compiler technology. After analyzing the code, we can get theknowledge of this program structure. It uses regular language, automata, context-free grammarand logic syntax tree said the program analysis process. The second step is to abstract theimportant information from the knowledge, design information abstraction rules and knowledgeorganization structure. After abstracting the knowledge, we can get the abstract information. Thethird step is to reconstruct software structure and other important information. It uses computervisualization technology, such as UML and Navier-Stokes graph. The method resolves theproblems of graphical design, graphical zoom, graphical localization, graphical nested and so on.Experimental results have shown that it can generate N-S figure quickly and effectively, thushelp the developers better understand the program source code efficiently. This paper lists theanalysis results of program comprehension system. It shows the table of database file path, thetable of key words, the table of classes, the table of function methods and the UMLgeneralization relationship between classes, N-S flow chart of function methods. At last, the paper analyzes the efficiency and accuracy of the system, which indicates that this theory isfeasible.