Automated detection of game loop implementation by the nest depth and number of iterations

Abstract

Game loops in the existing source code must be identified prior to updating, adding, or modifying game rules or game objects in games software. This paper proposes a static analysis method predicting game loop candidates with likelihood ranks by focusing on the characteristics of game loops. The proposed method is evaluated with ten publicly available games in Github. For eight of the games, actual game loops were predicted: five had an actual game loop listed at the top of the candidates, while the other three had an actual game loop within the top 11 results.