SAIFU: Supporting Program Understanding by Automatic Indexing of Functionalities in Source Code

• Main Authors: Masashi Nishimoto, Keiji Nishiyama, Hideyuki Kawabata, Tetsuo Hironaka
• Format: Article
• Language: English
• Published: Atlantis Press 2019-09-01
• Series: International Journal of Networked and Distributed Computing (IJNDC)
• Subjects: Program understanding; dataflow graph; tag cloud; event-driven programming
• Online Access: https://www.atlantis-press.com/article/125918015/view

Programs in the event-driven style that are typical of mobile and/or Web applications are becoming complex and hard to maintain. For the purpose of reducing the burden put on software developers while reading source code to understand its details, we propose a tool for supporting program understanding, named SAIFU (a tool for Supporting program understanding by Automatic Indexing of Functionalities). SAIFU automatically extracts implemented functionalities from source code and puts annotations to them. SAIFU helps the user grasp the behavior and the structure of a whole program by showing a list of the annotations of functionalities. SAIFU highlights a set of statements of the source code that are related to any functionality on the annotation list so that the user can investigate the implementation details of a particular functionality. Experimental results obtained by applying SAIFU to 16 applications in Google Samples confirm that the tool is effective for finding out important statements from existing Android application programs.