Semantic-based extraction approach for generating source code summary towards program comprehension

Abstract

Program comprehension is a vital process that involves much effort in software maintenance. A key challenge for the developers is to comprehend a software system to be maintained since it is difficult and time consuming. Nowadays, software systems have grown in size causing the increased of developers' tasks in exploring and understanding source code. Source code is a crucial resource for developers to become familiar with a software system since some system documentation is often unavailable or outdated. Although many researchers have discussed different strategies and techniques to overcome the program comprehension problem, there are still many challenges that they have not yet discovered when trying to understand a software system through reading source code. Therefore, this study attempts to overcome the problem of source code comprehension by suggesting a suitable comprehension technique. We propose a semantic-based extraction approach of source code and generating it as a summary. This work aims to explain the concepts and relationships of program by integrating utilization of ontology and UML class-based modeling approaches. It will be easier for maintainer to understand source code as well as create a better way for improving program comprehension.