Exploring Software Quality Through Data-Driven Approaches and Knowledge Graphs

Context: The quality of software systems has always been a crucial task and has led to the establishment of various reputable software quality models. However, the automation trends in Software Engineering have challenged the traditional notion of quality assurance, motivating the development of a new paradigm with advanced AI-based quality standards. Objective: The goal of this paper is to bridge the gap between theoretical frameworks and practical implementations on the aspects of software quality. Methodology: This study involved an extensive literature review of software quality models, including McCall, Boehm, Dromey, FURPS, and ISO/IEC 25010. The detailed information about quality attributes from each model was systematically synthesized and organized into datasets, data frames, and Python dictionaries. The resulting resources were then shared and made accessible through a public GitHub repository. Results: In brief, this research provides (i) a comprehensive dataset on software quality containing catalogs of quality models and attributes, (ii) a Python dictionary encapsulating the quality models and their associated characteristics for convenient empirical experimentation, (iii) the application of advanced knowledge graph techniques for the analysis and visualization of software quality parameters, and (iv) the complete construction steps and resources for download, ensuring easy integration and accessibility. Conclusion: This study builds a foundational step towards the standardization of automating software quality modeling to enhance not just quality but also efficiency for software development. For our future work, there will be a concentration on the practical utilization of the dataset in real-world software development contexts.