Research On Program Change Impact Analysis Based On Code Text Analysis

Nowadays, computer technology is developing rapidly and covers various areas, and has a far-reaching impact in people’s life. Software industry is a major branch of the computer industry, but the increasing social needs and the rapid development of computer industry give an important test to the development of software technology, meanwhile social development causes people’s incresing needs for software usability, aesthetics and other aspects of requirements. The development of software technology or other innovative models can improve the efficiency and propercy of the software development and maintenance in some degree. But with the expanding software scale and more complex business requirements, software maintainability has been reduced and maintenance costs have been raised, which caused a great loss to software developers and usres.Software change impact analysis(CIA) is a technology which is frequently used in the phase of development and maintenance. In a large software system, software change has a ripple effect, which results in a number of other changes. That is so harmfull for software’s development and maintenance. And ripple effect can causes software development and maintenance personnel spending a lot of time and energy to find the location of the fault. But change impact analysis can help software developers to find the code affected by the change automatically, and the software developers can modify it. Among a large amount of change impact analysis technologies, there are many technologies have been uesd as a tool. Due to change impact analysis involving text analysis, maybe we can apply other technologies to solve the CIA problem.Based on information retrieval and cloning code detection, we propose two change impact analysis technologies respectively, which are word assimilation technology and word differentiation technology. Application of information retrieval technology to change impact analysis is a typical method; this paper will also introduce the cloning code detection technology for change impact analysis innovatively. To analyze the most important factor for change impact analysis, in the experimental process, we compare the results of two technologies and analyze each one’s advantages and disadvantages between them, also we consider the impact of all the weights and threshold in each technology experiment. Through comparative analysis of the experiment, we draw a conclusion that the recall rate which calculated through word assimilation technology is higher than that calculated through word differentation technology. In other words, for the first technology, the precision of calculation results is worse than that of the second technology. So by overall consideration, the word diffrentation technology is better.