An eye tracking perspective on harmfulness of code smells: a systematic literature review and experiment design

Abstract

This work aims to update a prior systematic review focusing on the application of eye-tracking methodologies in software engineering, particularly concerning code smells. Code smells are features within source code that are difficult to read, understand, and maintain, yet their direct harmfulness remains insufficiently explored. Through a comprehensive literature review, 28 papers were identified showcasing recent advancements in eye-tracking studies from 2015 onward, highlighting new insights and limitations. It was observed that eye-trackers are used to study model comprehension, code comprehension, debugging, and traceability tasks. Based on these insights, we propose an experimental design employing eye-tracking methodology to analyze the harmful effects of code smells.