Studies on software fault-proneness were typically focused on analysis and prediction, with a few exceptions that used explanatory approach. This paper proposes a novel methodology that, for the first time, utilizes a case-control approach for building explanatory models of software fault-proneness. The files with post-release faults are treated as cases and the other files as controls. The cases and controls are matched by size and prerelease fault-proneness (i.e., Bugfixes) is treated as an exposure. The methodology incorporates software metrics as confounders and, for the first time, considers their interactions. Furthermore, the methodology rigorously handles multicollinearity and uses backward elimination to produce the simplest explanatory models. The odds ratios are quantified using conditional logistic regression which leads to efficient estimates, with tighter confidence intervals. The empirical results, based on three Eclipse releases, showed that while some metrics (i.e., Age, Bugfixes and Developers) consistently affected post-release fault-proneness in two or three releases, the effects of other metrics and interactions were release-specific. Additionally, the first-ever systematic exploration of the generalizability in prior explanatory studies showed that, similarly to our study, they experienced limited generalizability of impactful factors, which is likely due to the complex nature of the software and its development processes. Our results have several practical implications: (1) Simple models with 4–7 significant metrics and interactions can explain post-release fault-proneness; (2) The impactful metrics are simple and easy to collect (e.g., files age, the existence of prerelease faults, and developers count); (3) Due to limited generalizability, release/project-specific explanatory models are necessary.

Higher educational institutions seek to improve the quality and the productivity of the educational process. The current attitude is toward involving students in the learning and evaluation process. Peer review has been proven to be one of the most effective tactics to attain this in software engineering disciplines wherein project-based courses are substantial to afford high quality competencies. A few studies in literature empirically studied the impact of peer review in project-based software engineering courses. In this work, we attempt to provide more insights by implementing peer assessment in one of the project-based courses offered by the Department of Software Engineering at the Hashemite University, Jordan, that is “Object-Oriented Software Development”. In this paper, we investigate the validity of peer assessment by examining how well students in this course evaluate their peers and how the strength of students affects their assessment. The work also embeds a rubric that comprises key criteria to assess software system modelling. The results of the study reveal promising signs of using peer assessment in project-based software engineering courses.

Cybersecurity is the top priority for most businesses and government agencies. Teaching students practical cybersecurity is one of the challenges for most academic institutions. Some institutions focus on theory and neglect the practical, which is more appealing to students. Besides, it equips students with the skills the industry needs and reduces the gap between academic institutions and the industry. National Cyber League NCL provides a platform where all students nationwide can join and practice. In the fall of 2023, West Virginia University cybersecurity students joined the NCL competition as part of the course requirements. Participating in team and individual competitions was considered a substitute for the project for the course, which consists of 30% of the total grade. In this paper, we evaluated their background and skills before entering the competition (Who they are?), their experience in the competition, their achievements and skills after completing the competition (what they learned?), and how they evaluate their experience (what they think?). In this paper, we measure their involvement with the competition and how the competition assists them in developing their skills. We rely on their opinion and compare that with their performance according to the report card produced by NCL for every team and each individual. We also found that most students were involved between 10 to 30 hours during the semester with the games. Few students were more than 30 hours during the semester involved with NCL. We found that students were reasonably engaged in the competitions in different areas. Students spent more time in the gymnasium and practice games than in actual competitions. Students achieved an average completion of 35% with 73% average accuracy in the gymnasium, average completion was 19%, and the average accuracy was 76% in the practice game, average completion was 29%, and the average accuracy was 60% in the individual games, and 38% average completion and 39% average accuracy in the team games. Most students believed their skill level had improved by at least one or more. The majority of students were satisfied with their learning experience at NCL.

Abstract At the beginning of the testing phase and before the deployment phase of a project's development cycle, we need to predict files with a high chance of change. Software products are always prone to change due to several reasons, including fixing errors or improvements. In this work, we used the Eclipse (releases from 2.0 to 3.5) to investigate how prediction models can perform when learning from a release and predicting in the subsequent one, which contains new files that models have not seen. We compared the performance of these models with models that are trained and tested on the same release. We found no differences between predicting the same release or subsequent release on two pre Europa releases. Predicting change in newly created files helps improve maintenance planning for software project managers and reduce cost. It will also help to enhance the quality of software by improving the practices of developers. This study used the Adaptive Boost classifier with the decision tree J48 algorithm and combined it with the re‐sampling method. We find this to be better than using a meta classifier alone or combine the re‐sampling with the standard classification. We compared our results with related works and found that our results are outperforming.

Machine learning algorithms have been used extensively for software fault proneness prediction. This paper presents the first application of Group Lasso Regression (G-Lasso) for software fault proneness classification and compares its performance to six widely used machine learning algorithms. Furthermore, we explore the effects of two factors on the prediction performance: the effect of imbalance treatment using the Synthetic Minority Over-sampling Technique (SMOTE), and the effect of datasets used in building the prediction models. Our experimental results are based on 22 datasets extracted from open source projects. The main findings include: (1) G-Lasso is robust to imbalanced data and significantly outperforms the other machine learning algorithms with respect to the Recall and G-Score, i.e., the harmonic mean of Recall and (1- False Positive Rate). (2) Even though SMOTE improved the performance of all learners, it did not have statistically significant effect on G-Lasso's Recall and G-Score. Random Forest was in the top performing group of learners for all performance metrics, while Naive Bayes performed the worst of all learners. (3) When using the same change metrics as features, the choice of the dataset had no effect on the performance of most learners, including G-Lasso. Naive Bayes was the most affected, especially when balanced datasets were used.