An Empirical Investigation into Code Smell Elimination Sequences for Energy Efficient Software

摘要

Recent research has shown that maintainability improving activities, such as removing code smells using recommended refactoring activities, may degrade the energy consumption behavior of software systems. However, current research is still immature and requires considerable effort for transferring findings to practice. This work empirically investigates the impact of eliminating a set of three notorious code smells, individually as well as in all six possible sequences, on energy consumption behavior of software systems. It also analyzes whether any relationship exists between software architecture sustainability (in terms of energy efficiency) and maintainability within the context of individual and sequential code smell elimination. The study outcomes show that the selected code smell removal permutations yield variant levels of energy consumption values for the resulted refactored software versions. Also, a particular permutation is learned to yield most energy-efficient refactored software versions, when compared to all other code smell removal permutations.