Visualizing code bad smells

Abstract

Software visualization is an effective way to support human comprehension to large software systems. In software maintenance, most of the time is spent on understanding code in order to change it. This paper presents a visualization approach to help maintainers to locate and understand code bad smells. Software maintainers need to locate and understand these bad smells in order to remove them via code refactoring. Object oriented code elements are visualized as well as their bad smells if they exist. The proposed visualization shows classes as building and bad smell as letter avatars based on the initials of the names of bad smells. These avatars are shown as warning signs on the buildings. A framework is proposed to automatically analyze code to identify bad smells and to generate the proposed visualizations. The evaluation of the proposed visualizations showed they reduce the comprehension time needed to understand bad smells.