| 1. |
- Barney, Sebastian, et al.
(författare)
-
Improving Students With Rubric-Based Self-Assessment and Oral Feedback
- 2012
-
Ingår i: IEEE Transactions on Education. - : IEEE. - 0018-9359. ; 55:3, s. 319-325
-
Tidskriftsartikel (refereegranskat)abstract
- Rubrics and oral feedback are approaches to help students improve performance and meet learning outcomes. However, their effect on the actual improvement achieved is inconclusive. This paper evaluates the effect of rubrics and oral feedback on student learning outcomes. An experiment was conducted in a software engineering course on requirements engineering, using the two approaches in course assignments. Both approaches led to statistically significant improvements, though no material improvement (i.e., a change by more than one grade) was achieved. The rubrics led to a significant decrease in the number of complaints and questions regarding grades.
|
|
| 2. |
- Borg, Markus, et al.
(författare)
-
Test automation with grad-CAM Heatmaps - A future pipe segment in MLOps for Vision AI?
- 2021
-
Ingår i: Proceedings - 2021 IEEE 14th International Conference on Software Testing, Verification and Validation Workshops, ICSTW 2021. - : Institute of Electrical and Electronics Engineers Inc.. - 9781665444569 ; , s. 175-181
-
Konferensbidrag (refereegranskat)abstract
- Machine Learning (ML) is a fundamental part of modern perception systems. In the last decade, the performance of computer vision using trained deep neural networks has outperformed previous approaches based on careful feature engineering. However, the opaqueness of large ML models is a substantial impediment for critical applications such as in the automotive context. As a remedy, Gradient-weighted Class Activation Mapping (Grad-CAM) has been proposed to provide visual explanations of model internals. In this paper, we demonstrate how Grad-CAM heatmaps can be used to increase the explainability of an image recognition model trained for a pedestrian underpass. We argue how the heatmaps support compliance to the EU's seven key requirements for Trustworthy AI. Finally, we propose adding automated heatmap analysis as a pipe segment in an MLOps pipeline. We believe that such a building block can be used to automatically detect if a trained ML-model is activated based on invalid pixels in test images, suggesting biased models.
|
|
| 3. |
- Duc, Anhnguyen, et al.
(författare)
-
Security challenges in IoT development : A software engineering perspective
- 2017
-
Ingår i: ACM International Conference Proceeding Series. - New York, NY, USA : Association for Computing Machinery (ACM). - 9781450352642
-
Konferensbidrag (refereegranskat)abstract
- The rapid growth of Internet-of-things (IoT) software applications has driven both practitioners and researchers' attention to methodological approaches for secure IoT development. Security issues for IoT is special in the way that they include not only software, but also hardware and network concerns. With the aim at proposing a methodological approach for secure IoT application development, we investigated what are security challenges in the context of IoT development. We reviewed literature and investigated two industry cases. The preliminary finding results in a list of 17 security challenges with regards to technical, organizational and methodological perspectives. Cross-case comparison provides initial explanation about the less emphasis on methodological and organizational security concerns in our cases. © 2017 ACM.
|
|
| 4. |
- Jabangwe, Ronald, et al.
(författare)
-
A method for investigating the quality of evolving object-oriented software using defects in global software development projects
- 2016
-
Ingår i: Journal of Software. - : John Wiley & Sons. - 2047-7473 .- 2047-7481. ; 28:8, s. 622-641
-
Tidskriftsartikel (refereegranskat)abstract
- Context: Global software development (GSD) projects can have distributed teams that work independently in different locations or team members that are dispersed. The various development settings in GSD can influence quality during product evolution. When evaluating quality using defects as a proxy, the development settings have to be taken into consideration. Objective: The aim is to provide a systematic method for supporting investigations of the implication of GSD contexts on defect data as a proxy for quality. Method: A method engineering approach was used to incrementally develop the proposed method. This was done through applying the method in multiple industrial contexts and then using lessons learned to refine and improve the method after application. Results: A measurement instrument and visualization was proposed incorporating an understanding of the release history and understanding of GSD contexts. Conclusion: The method can help with making accurate inferences about development settings because it includes details on collecting and aggregating data at a level that matches the development setting in a GSD context and involves practitioners at various phases of the investigation. Finally, the information that is produced from following the method can help practitioners make informed decisions when planning to develop software in comparable circumstances. Copyright © 2016 John Wiley & Sons, Ltd. Copyright © 2016 John Wiley & Sons, Ltd.
|
|
| 5. |
- Jabangwe, Ronald, et al.
(författare)
-
An Exploratory Study of Software Evolution and Quality : Before, During and After a Transfer
- 2012
-
Konferensbidrag (refereegranskat)abstract
- In the light of globalization it is not uncommon that different teams from different locations get involved in the development of a software product during its evolution. However, empirical evidence that demonstrates the effect of changing team members on software quality is scarce. In this paper, we investigate quality of a software product, across subsequent software releases, that was first developed in one location of Ericsson, a large multinational corporation, then jointly with an offshore location of the same company, and finally transferred offshore. To get a better understanding multiple data sources are used in the analysis: qualitative data, consisting of interviews and documentation, and quantitative data, consisting of release history and defect statistics. Our findings confirm an initial decline in efficiency and quality after a transfer, and highlight the areas of concern for companies that are considering transferring their product development from experienced teams to those having limited or no previous engagement with the product.
|
|
| 6. |
- jabangwe, Ronald, et al.
(författare)
-
Decision support for offshore insourcing software development
- 2011
-
Konferensbidrag (refereegranskat)abstract
- Proximity to customers and the possibility of accessing new and prospective markets are some of the factors that compel companies to consider various sourcing options and go global. One sourcing option that has consequently become popular is Offshore insourcing. Though it is appealing as a business strategy due to the perceived benefits, there are some strategies that may inexplicably impact software quality. The objective of this research work is to focus on producing empirical evidence that is valuable input to support decision-making during offshore insourcing projects. The input will help with evaluating different strategies on their impact on quality.
|
|
| 7. |
- Jabangwe, Ronald, et al.
(författare)
-
Distributed Software Development in an Offshore Outsourcing Project : A Case Study of Source Code Evolution and Quality
- 2016
-
Ingår i: Information and Software Technology. - : Elsevier. - 0950-5849 .- 1873-6025. ; 72, s. 125-136
-
Tidskriftsartikel (refereegranskat)abstract
- Context: Offshore outsourcing collaborations can result in distributed development, which has been linked to quality-related concerns. However, there are few studies that focus on the implication of distributed development on quality, and they report inconsistent findings using different proxies for quality. Thus, there is a need for more studies, as well as to identify useful proxies for certain distributed contexts. The presented empirical study was performed in a context that involved offshore outsourcing vendors in a multisite distributed development setting.Objective: The aim of the study is to investigate how quality changes during evolution in a distributed development environment that incurs organizational changes in terms of number of companies involved.Method: A case study approach is followed in the investigation. Only post-release defects are used as a proxy for external quality due to unreliable defect data found pre-release such as those reported during integration. Focus group meetings were also held with practitioners.Results: The results suggest that practices that can be grouped into product, people, and process categories can help ensure post-release quality. However, post-release defects are insufficient for showing a conclusive impact on quality of the development setting. This is because the development teams worked independently as isolated distributed teams, and integration defects would help to better reflect on the impact on quality of the development setting.Conclusions: The mitigation practices identified can be useful information to practitioners that are planning to engage in similar globally distributed development projects. Finally, it is important to take into consideration the arrangement of distributed development teams in global projects, and to use the context to identify appropriate proxies for quality in order to draw correct conclusions about the implications of the context. This would help with providing practitioners with well-founded findings about the impact on quality of globally distributed development settings.
|
|
| 8. |
- Jabangwe, Ronald, et al.
(författare)
-
Empirical Evidence on the Link between Object-Oriented Measures and External Quality Attributes : A Systematic Literature Review
- 2015
-
Ingår i: Empirical Software Engineering. - : Springer. - 1382-3256 .- 1573-7616. ; 20:3, s. 640-693
-
Tidskriftsartikel (refereegranskat)abstract
- There is a plethora of studies investigating object-oriented measures and their link with external quality attributes, but usefulness of the measures may differ across empirical studies. This study aims to aggregate and identify useful object-oriented measures, specifically those obtainable from the source code of object-oriented systems that have gone through such empirical evaluation. By conducting a systematic literature review, 99 primary studies were identified and traced to four external quality attributes: reliability, maintainability, effectiveness and functionality. A vote-counting approach was used to investigate the link be- tween object-oriented measures and the attributes, and to also assess the consistency of the relation reported across empirical studies. Most of the studies investigate links between object-oriented measures and proxies for reliability attributes, followed by proxies for maintainability. The least investigated attributes were: effectiveness and functionality. Measures from the C&K measurement suite were the most popular across studies. Vote-counting results suggest that complexity, cohesion, size and coupling measures have a better link with reliability and maintainability than inheritance measures. However, inheritance measures should not be overlooked during quality assessment initiatives; their link with reliability and maintainability could be context dependent. There were too few studies traced to effectiveness and functionality attributes; thus a meaningful vote-counting analysis could not be conducted for these attributes. Thus, there is a need for diversification of quality attributes investigated in empirical studies. This would help with identifying useful measures during quality assessment initiatives, and not just for reliability and maintainability aspects.
|
|
| 9. |
- Jabangwe, Ronald, et al.
(författare)
-
Handover of managerial responsibilities in global software development : a case study of source code evolution and quality
- 2015
-
Ingår i: Software quality journal. - : Springer. - 0963-9314 .- 1573-1367. ; 23:4, s. 539-566
-
Tidskriftsartikel (refereegranskat)abstract
- Studies report on the negative effect on quality in global software development (GSD) due to communication and coordination-related challenges. However, empirical studies reporting on the magnitude of the effect are scarce. This paper presents findings from an embedded explanatory case study on the change in quality over time, across multiple releases, for products that were developed in a GSD setting. The GSD setting involved periods of distributed development between geographically dispersed sites as well as a handover of project management responsibilities between the involved sites. Investigations were performed on two medium-sized products from a company that is part of a large multinational corporation. Quality is investigated quantitatively using defect data and measures that quantify two source code properties, size and complexity. Observations were triangulated with subjective views from company representatives. There were no observable indications that the distribution of work or handover of project management responsibilities had an impact on quality on both products. Among the product-, process- and people-related success factors, we identified well-designed product architectures, early handover planning and support from the sending site to the receiving site after the handover and skilled employees at the involved sites. Overall, these results can be useful input for decision-makers who are considering distributing development work between globally dispersed sites or handing over project management responsibilities from one site to another. Moreover, our study shows that analyzing the evolution of size and complexity properties of a product’s source code can provide valuable information to support decision-making during similar projects. Finally, the strategy used by the company to relocate responsibilities can also be considered as an alternative for software transfers, which have been linked with a decline in efficiency, productivity and quality.
|
|
| 10. |
- Jabangwe, Ronald
(författare)
-
Software Quality Evaluation for Evolving Systems in Distributed Development Environments
- 2015
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Context: There is an overwhelming prevalence of companies developing software in global software development (GSD) contexts. The existing body of knowledge, however, falls short of providing comprehensive empirical evidence on the implication of GSD contexts on software quality for evolving software systems. Therefore there is limited evidence to support practitioners that need to make informed decisions about ongoing or future GSD projects. Objective: This thesis work seeks to explore changes in quality, as well as to gather confounding factors that influence quality, for software systems that evolve in GSD contexts. Method: The research work in this thesis includes empirical work that was performed through exploratory case studies. This involved analysis of quantitative data consisting of defects as an indicator for quality, and measures that capture software evolution, and qualitative data from company documentations, interviews, focus group meetings, and questionnaires. An extensive literature review was also performed to gather information that was used to support the empirical investigations. Results: Offshoring software development work, to a location that has employees with limited or no prior experience with the software product, as observed in software transfers, can have a negative impact on quality. Engaging in long periods of distributed development with an offshore site and eventually handing over all responsibilities to the offshore site can be an alternative to software transfers. This approach can alleviate a negative effect on quality. Finally, the studies highlight the importance of taking into account the GSD context when investigating quality for software that is developed in globally distributed environments. This helps with making valid inferences about the development settings in GSD projects in relation to quality. Conclusion: The empirical work presented in this thesis can be useful input for practitioners that are planning to develop software in globally distributed environments. For example, the insights on confounding factors or mitigation practices that are linked to quality in the empirical studies can be used as input to support decision-making processes when planning similar GSD projects. Consequently, lessons learned from the empirical investigations were used to formulate a method, GSD-QuID, for investigating quality using defects for evolving systems. The method is expected to help researchers avoid making incorrect inferences about the implications of GSD contexts on quality for evolving software systems, when using defects as a quality indicator. This in turn will benefit practitioners that need the information to make informed decisions for software that is developed in similar circumstances.
|
|
