How Aware Are We of Test Smells in Quantum Software Systems? A Preliminary Empirical Evaluation

• Tássio Virgínio UFBA
• Larissa Bastos UNEB
• Carla Bezerra UFC
• Márcio Ribeiro UFAL
• Ivan Machado UFBA

Resumo

Context: With the rapid progress of quantum computing, Quantum Software Engineering (QSE) is establishing itself as an essential discipline to support developers throughout all stages of quantum software development. The area of testing in quantum systems has received greater attention in research on this topic to guarantee the quality and reliability of these technologies. Objective: This paper presents an empirical study focused on the testing of quantum software at its classical layer. Specifically, it aims to identify and analyze the unique characteristics of quantum software tests, particularly in terms of Test Smells, their distribution, recurrence, and differences compared to classical software tests. Method: We used two sets of software from previous studies, one comprising 12 quantum software and the other comprising 80 classical software. From these datasets, we conducted an analysis to detect 10 test smells, allowing us to map their dispersion in quantum software, identify their specific characteristics, and draw comparisons with classical software. Results: Our findings reveal a high dispersion of test smells of 51% in quantum software. Furthermore, quantum tests exhibit statistical differences from classical software tests, with the most outlier being Conditional Test Logic, which is 20% more frequent than in classical software. Conclusions: The insights gained from this study can contribute to enhancing the quality, maintainability, and readability of tests written for the classical layer of quantum software. Ultimately, this can improve the overall understanding and quality of quantum software.

Palavras-chave: Quantum Software Systems, Software Testing, Test Smells, Empirical Evaluation