The Impact of Code Smells on Energy Consumption: A study about Model-View-Controller Applications
收藏资源简介:
Software engineering recognizes the importance of quality attributes. However, software sustainability—particularly in terms of energy consumption—has received comparatively less attention. Understanding how software quality influences energy consumption is essential to support informed decisions toward more sustainable software development. This work investigates the relationship between software quality and energy consumption, with a particular focus on maintainability, a key quality attribute closely associated with code smells and technical debt. The study analyzes how maintainability-related issues affect energy consumption and how this impact is influenced by software size and functionality type. To isolate the effect of maintainability while keeping technical characteristics constant, the study focuses on Java desktop applications following the Model–View–Controller (MVC) architectural pattern. An exploratory study was conducted on 44 applications, consisting of 22 original applications and their corresponding 22 altered versions, in which maintainability-related issues were systematically introduced. The analysis considers four representative code smells—unused variables, long parameter lists, duplicated code, and duplicated files—along with technical debt. This Zenodo repository contains the datasets supporting this study. The 132 energy consumption measurements obtained from the 44 applications are provided in the compressed file Applications Energy Consumption Measurements.zip. Additionally, the synthesized and refined dataset used for the statistical analysis is available in the file VApp_Data_Refined.xlsx, which integrates energy consumption data with maintainability metrics and application size. These datasets enable replication of the study and further research on the impact of maintainability on the energy efficiency and sustainability of Java desktop MVC applications.



