Journal of Software-Evolution and Process最新文献

Evolutionary test case generation plays a vital role in ensuring software quality and reliability. Since Web applications involve a large number of interactions between client and server, the dynamic evolutionary test case generation is very time-consuming, which makes it difficult to apply in actual projects. Obviously, parallelization provides a feasible way to improve the efficiency and effectiveness of evolutionary test generation. In our previous research, the idea of parallelism has been introduced into the evolutionary test generation for Web applications. However, its performance is affected by many factors, such as migration scale, migration frequency, the number of browser processes and subpopulations, and so on. The analysis of influencing factors can guide enhancing the performance of evolutionary test generation. For this reason, this paper analyzes the factors that influence parallel evolutionary algorithms and how they affect the performance of test generation for Web applications. At the same time, different parallel evolutionary test generation methods are designed and implemented. Experiments are conducted on open-source Web applications to generate test cases that meet the server-side sensitive paths coverage criterion, providing guidance and suggestions for the parameter setting of parallel evolutionary test case generation for Web applications. The experimental results show that (1) compared with the global parallelization model, the evolutionary algorithm based on the parallel island model has a greater improvement in test case generation performance. In more detail, when generating test cases with the same server-side sensitive paths coverage, the number of iterations required is reduced by 49.6%, and the time cost is reduced by 58.7%; (2) for the test case generation based on the parallel island model, if the migration scale is large, appropriately increasing the migration frequency can reduce its time cost; (3) if the number of subpopulations is fixed, appropriately increasing the number of browser processes can reduce the time cost of Web application test case evolution, but the number of browser processes should not be too large; otherwise, it may increase the time cost.

Patches can help fix security vulnerabilities and optimize software performance, thereby enhancing the quality and security of the software. Unfortunately, patches generated by automated program repair tools are not always correct, as they may introduce new bugs or fail to fully rectify the original issue. Various methods for evaluating patch correctness have been proposed. However, most methods face the challenge of capturing long-distance dependencies in patch correctness evaluation, which leads to a decline in the predictive performance of the models. To address the challenge, this paper presents a method named Qamhaen to evaluate the correctness of patches generated by APR. Specifically, text embedding of bugs and patches component address the challenge of long-distance dependencies across functions in patch correctness evaluation by using bug reports and patch descriptions as inputs instead of code snippets. BERT is employed for pretraining to capture these dependencies, followed by an additional multihead self-attention mechanism for further feature extraction. Similarity evaluator component devises a similarity calculation to assess the effectiveness of patch descriptions in resolving issues outlined in bug reports. Comprehensive experiments are conducted on a dataset containing 9135 patches and a patch correctness assessment metric, and extensive experiments demonstrate that Qamhaen outperforms baseline methods in terms of overall performance across AUC, F1, +Recall, -Recall, and Precision. For example, compared to the baseline, Qamhaen achieves an F1 of 0.691, representing improvements of 24.2%, 22.1%, and 6.3% over the baseline methods, respectively.

As software development grows increasingly complex, application programming interface (API) plays a significant role in enhancing development efficiency and code quality. However, the explosive growth in the number of APIs makes it impossible for developers to become familiar with all of them. In actual development scenarios, developers may spend a significant amount of time searching for suitable APIs, which could severely impact the development process. Recently, the OpenAI's large language model (LLM) based application—ChatGPT has shown exceptional performance across various software development tasks, responding swiftly to instructions and generating high-quality textual responses, suggesting its potential in API recommendation tasks. Thus, this paper presents an empirical study to investigate the performance of ChatGPT in query-based API recommendation tasks. Specifically, we utilized the existing benchmark APIBENCH-Q and the newly constructed dataset as evaluation datasets, selecting the state-of-the-art models BIKER and MULAREC for comparison with ChatGPT. Our research findings demonstrate that ChatGPT outperforms existing approaches in terms of success rate, mean reciprocal rank (MRR), and mean average precision (MAP). Through a manual examination of samples in which ChatGPT exceeds baseline performance and those where it provides incorrect answers, we further substantiate ChatGPT's advantages over the baselines and identify several issues contributing to its suboptimal performance. To address these issues and enhance ChatGPT's recommendation capabilities, we employed two strategies: (1) utilizing a more advanced LLM (GPT-4) and (2) exploring a new approach—MACAR, which is based on the Chain of Thought methodology. The results indicate that both strategies are effective.

Modern supply chains of goods and services rely heavily on close collaborations between the partners within these supply chains. Consequently, there is a demand for IT systems that support collaborations between business partners, for instance, allowing for joint computations for global optimizations (in contrast to local optimizations that each partner can do on their own). Still, businesses are very reluctant to share data or connect their enterprise systems to allow for such joint computation. The topmost factor that businesses name as reason for not collaborating, is their security concern in general and, in particular, the confidentiality of business critical data. While there are techniques (e.g., homomorphic encryption or secure multiparty computation) that allow joint computations and, at the same time, that are protecting the confidentiality of the data that flows into such a joint computation, they are not widely used. One of the main problems that prevent their adoption is their perceived performance overhead. In this paper, we address this problem by an approach that utilized the structure of supply chains by decomposing global computations into local groups, and applying secure multiparty computation within each group. This results in a scalable (resulting in a significant smaller runtime overhead than traditional approaches) and secure (i.e., protecting the confidentiality of data provided by supply chain partners) approach for joint computations within supply chains. We evaluate our approach using life-cycle assessment (LCA) as a case study. Our experiments show that, for instance, secure LCA computations even in supply chains with 15 partners are possible within less than two minutes, while traditional approaches using secure multiparty computation need more than a day.

Business process management (BPM) combined with new technologies can trigger both incremental and disruptive improvements in how organizations operate. More specifically, today's fourth industrial revolution can bring rapid changes in an organization's process dynamics. Our study explores differences between possible process-technology “fit” and “unfit” situations in BPM innovative projects. We extend relevant past studies and theories using a mix of qualitative techniques consisting of expert panel interviews and a case design using two field studies. Our findings reveal that, although alternative process-technology “fit” and “no-fit” situations exist, elements such as creativity, efficiency, integration, user friendliness, and proper task monitoring turn out to be the most promising factors to gain a process-technology fit. Novelty in our work includes discovering “fit” and “no-fit” factors in terms of process-technology alignment, and the development of a decision framework with a generic set of suggestions for BPM practitioners and decision makers. Our mixed-method approach is based on qualitative results by emphasizing in-depth insights and lessons learned rather than building a generalizable theory. We intend to guide managers and decision makers to help them think about possible directions, as suggested by our experts and case participants at the time of their technology adoption in a BPM context.

Understanding the evolution of software systems is crucial for advancing software engineering practices. Many studies have been devoted to exploring software evolution. However, they primarily treat software as an entire entity and overlook the inherent differences between subsystems, which may lead to biased conclusions. In this study, we attempt to explore variations between subsystems by investigating the code feature differences between kernel and nonkernel components from an evolutionary perspective. Based on three operating systems as case studies, we examine multiple dimensions, including the code churn characteristics and code inherent characteristics. The main findings are as follows: (1) The proportion of kernel code remains relatively small, and exhibits consistent stability across the majority of versions as systems evolve. (2) Kernel code exhibits higher stability in contrast to nonkernel code, characterized by a lower modification rate and finer modification granularity. The patterns of modification activities are similar in both kernel and nonkernel code, with a preference of changing code and a tendency to avoid the combination of adding and deleting code. (3) The cumulative code size and complexity of kernel files show an upward trajectory as the system evolves. (4) Kernel files exhibit a significantly higher code density and complexity than nonkernel files, featuring a greater number of code line, comments, and statements, along with a larger program length, vocabulary, and volume. Conversely, kernel functions prioritize modularity and maintainability, with a significantly smaller size and lower complexity than nonkernel functions. These insights contribute to a deeper understanding of the dynamics within operating system codebases and highlight the necessity of targeted maintenance strategies for different subsystems.