Feature Location Of Object-Oriented Software

More and more information society cannot work regularly without the support of software systems. So competent understanding of the system is one of the critical factors for the maintenance of the software system. The maintenance engineers should analyze the target system and get the information and knowledge they need, so as to finish different tasks. But program comprehension is so complex and costly, especially for the large and complicated systems, that prompt and accurate understanding of the system cannot be achieved, which makes the maintenance a failure.Many studies show that effective understanding of the software system should be based on the pertinent and detailed comprehension of maintenance task. This paper, based on the research on the 0-0 software and the current technologies of this area, gives out a systematic approach to feature location in object-oriented program. The core study contents contain the following aspects.Researches have been done based on the current approaches of feature location, which contains two methods of feature location as static structure and program dynamic system. In order to get accurate features and the traceability between features and code artifacts, IAFLO has been put out on the target of 0-0 software. This approach makes use of the reconnaissance technipues and establishes tracking relations between the features and codes artifacts. And then, make these artifacts as seeds to design concerned graphs for the 0-0 program, and establish the relationship of calling, reading, writing, and creating among classes, fields and methods. Concern graphs can only locate the approach of the classes and the level of fields. At this time, if it is needed to track the codes of methods, IAFLO structure applies the slicing approach of 0-0 software, by the standard of program slicing for the concerned fields and related language; it then calculates the program slice and finally gets the languages and codes which have direct effect on the feature realization. This approach can track the related codes artifacts swiftly and accurately from the external features of 0-0 software, so as to improve the efficiency of program understanding for the purpose of software maintenance.In this paper, the Integrated Feature Location Supporting Tool (IFLST) has been designed from the approaches that IAFLO has developed. The tool can locate features in the target program, and give concentrated and systematic support.