Research On Supporting Multi-language Understanding And Information Extraction Technology For Program Understanding

Starting from the birth of a variety of software systems, software maintenance, updating and upgrading accompanied the whole life cycle of the software. Software maintenance, updating and upgrading of the original system have high cost of manpower and material resources for the process analyzed and understood. With the development of program understanding technology, there are several advantages for software maintenance, updating and upgrading. First, the provided support and assistance are non-negligible. Second, the efficiency is improved for understanding the original programmer system. Moreover, the cost of software maintenance, updating and upgrading is reduced. Currently, a variety of programming languages are used for software. So a program understanding system, supporting multiple programming languages, will effectively improve the program understanding efficiency and reduce the cost of development.Program understanding is a class of technology that obtains a variety of users required information source by analyzing and information extraction, such as class diagrams, system architecture, documents, etc. The essential difference for different programming language reflects between lexical rules and grammar rules, while the method of program understanding is similar. The lexical rules and grammar rules of programming language is expressed by production. By designing a production-based multilingual program information extractor, the different programming languages achieve a uniform manner. Then, a program understanding system can be used to support multiple programming languages.The designed multi-language program understanding system of production-based includes three modules: scanning analyzer module, knowledge management module and results expressed module. The function of scanner in scanning analyzer module is to realize the recognition of different programming languages. The parser achieves the input source code analysis and information extraction by calling the designed production-based multilingual program information extractor. Knowledge management module consists of information database, information classification and storage of information structure. The function of knowledge management module is to classify and structural store the extracted information. By using the knowledge management module, the extracted original program information of multiple programming languages can be managed in a uniform form. The function of results expressed module is to show the final results of the program understanding to user in a convenient way. The results expressed module includes the results of UML class diagrams expressed and the results of XML document expressed. The project file of original program is inputted system by using scanning analyzer, and its contained information is displayed to user by UML class diagrams and XML document for results expressed module.