Enhanced Software Testing Model Under Software Warranty Policy Considering Debugger Learning and Time Postponement Factors
Wang Li, Chih‐Chiang Fang · IEEE Access · 2025
In the software development life cycle, ensuring high-quality and reliable software is crucial for developers. Unreliable software can result in customer loss, decreased revenue, and heightened operational risks stemming from defective codes. It is crucial for software developers to engage in rigorous software testing to minimize potential defects. However, striving for perfect and error-free software is often impractical due to constraints in budget, time, and testing resources. This necessitates a balanced approach, where developers must find a middle ground between ensuring software reliability and managing the costs of testing. The proposed model enables software developers to explore various options for conducting a software testing project, each with its own unique distribution of human resources. This allows for the selection of the most effective option. The model takes into account the learning curve, which can significantly impact the efficiency of the software testing process. This study considers not only the human learning factor involved but also the time delay factor in the debugging phase. It proposes a software reliability growth model that calculates both the associated costs and a reliability indicator. Furthermore, this study broadens its scope by examining the categorization of errors, assessing their impact on the system, and estimating the time needed to rectify simple versus complex errors using different truncated exponential distributions. To demonstrate the practical application of this model, numerical examples are provided, along with sensitivity analyses. These analyses offer valuable managerial insights and guidance, aiding in the formulation of informed strategies for software releases.