Search

In the software development life cycle, the quality and reliability of software are critical to software developers. Poor quality and reliability not only cause the loss of customers and sales but also increase the operational risk due to unreliable codes. Therefore, software developers should try their best to reduce such potential software defects by undertaking a software testing project. However, to pursue perfect and faultless software is unrealistic since the budget, time, and testing resources are limited, and the software developers need to reach a compromise that balances software reliability and the testing cost. Using the model presented in this study, software developers can devise multiple alternatives for a software testing project, and each alternative has its distinct allocation of human resources. The best alternative can therefore be selected. Furthermore, the allocation incorporates debuggers' learning and negligent factors, both of which influence the efficiency of software testing in practice. Accordingly, the study considers both human factors and the nature of errors during the debugging process to develop a software reliability growth model to estimate the related costs and the reliability indicator. Additionally, the issue of error classification is also extended by considering the impacts of errors on the system, and the expected time required to remove simple or complex errors can be estimated based on different truncated exponential distributions. Finally, numerical examples are presented and sensitivity analyses are performed to provide managerial insights and useful directions to inform software release strategies. [ABSTRACT FROM AUTHOR]

In this study, an imperfect debugging software reliability growth model (SRGM) with Bayesian analysis was proposed to determine an optimal software release in order to minimize software testing costs and also enhance the practicability. Generally, it is not easy to estimate the model parameters by applying MLE (maximum likelihood estimation) or LSE (least squares estimation) with insufficient historical data. Therefore, in the situation of insufficient data, the proposed Bayesian method can adopt domain experts' prior judgments and utilize few software testing data to forecast the reliability and the cost to proceed with the prior analysis and the posterior analysis. Moreover, the debugging efficiency involves testing staff's learning and negligent factors, and therefore, the human factors and the nature of debugging process are taken into consideration in developing the fundamental model. Based on this, the estimation of the model's parameters would be more intuitive and can be easily evaluated by domain experts, which is the major advantage for extending the related applications in practice. Finally, numerical examples and sensitivity analyses are performed to provide managerial insights and useful directions for software release strategies. [ABSTRACT FROM AUTHOR]

Fault-counting data are collected in the testing process of software development. However, the data are not used for evaluating the efficiency of fault correction activities because the information on the fault detection and correction times of each fault are not recorded in the fault-counting data. Furthermore, it is difficult to collect new data on the detection time of each fault to realize efficiency evaluation for fault correction activities from the collected fault-counting data due to the cost of personnel and data collection. In this paper, we apply the thinning method, using intensity functions of the delayed S-shaped and inflection S-shaped software reliability growth models (SRGMs) to generate sample data of the fault detection time from the fault-counting data. Additionally, we perform simulations based on the infinite server queuing model, using the generated sample data of the fault detection time to visualize the efficiency of fault correction activities. [ABSTRACT FROM AUTHOR]

This paper presents a general testing coverage software reliability modeling framework that covers imperfect debugging and considers not only fault detection processes (FDP) but also fault correction processes (FCP). Numerous software reliability growth models have evaluated the reliability of software over the last few decades, but most of them attached importance to modeling the fault detection process rather than modeling the fault correction process. Previous studies analyzed the time dependency between the fault detection and correction processes and modeled the fault correction process as a delayed detection process with a random or deterministic time delay. We study the quantitative dependency between dual processes from the viewpoint of fault amount dependency instead of time dependency, then propose a generalized modeling framework along with imperfect debugging and testing coverage. New models are derived by adopting different testing coverage functions. We compared the performance of these proposed models with existing models under the context of two kinds of failure data, one of which only includes observations of faults detected, and the other includes not only fault detection but also fault correction data. Different parameter estimation methods and performance comparison criteria are presented according to the characteristics of different kinds of datasets. No matter what kind of data, the comparison results reveal that the proposed models generally give improved descriptive and predictive performance than existing models. [ABSTRACT FROM AUTHOR]

It is known that quantitative measures for the reliability of software systems can be derived from software reliability models, and, as such, support the product development process. Over the past four decades, research activities in this area have been performed. As a result, many software reliability models have been proposed. It was shown that, once these models reach a certain level of convergence, it can enable the developer to release the software and stop software testing accordingly. Criteria to determine the optimal testing time include the number of remaining errors, failure rate, reliability requirements, or total system cost. In this paper, we present our results in predicting the reliability of software for agile testing environments. We seek to model this way of working by extending the Jelinski–Moranda model to a "stack" of feature-specific models, assuming that the bugs are labeled with the features they belong to. In order to demonstrate the extended model, two use cases are presented. The questions to be answered in these two cases are: how many software bugs remain in the software and should one decide to stop testing the software? [ABSTRACT FROM AUTHOR]