| 1. |
- Knauss, Eric, 1977, et al.
(författare)
-
Supporting Continuous Integration by Code-Churn Based Test Selection
- 2015
-
Ingår i: RCoSE - 2nd International Workshop on Rapid Continuous Software Engineering @ ICSE 2015 Florence, Italy.
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Continuous integration promises advantages in large-scale software development by enabling software development organizations to deliver new functions faster. However, implementing continuous integration in large software development organizations is challenging because of organizational, social and technical reasons. One of the technical challenges is the ability to rapidly prioritize the test cases which can be executed quickly and trigger the most failures as early as possible. In our research we propose and evaluate a method for selecting a suitable set of functional regression tests on system level. The method is based on analysis of correlations between test-case failures and source code changes and is evaluated by combining semi-structured interviews and workshops with practitioners at Ericsson and Axis Communications in Sweden. The results show that using measures of precision and recall, the test cases can be prioritized. The prioritization leads to finding an optimal test suite to execute before the integration.
|
|
| 2. |
- Al Sabbagh, Khaled, 1987, et al.
(författare)
-
Predicting Test Case Verdicts Using TextualAnalysis of Commited Code Churns
- 2019
-
Ingår i: CEUR Workshop Proceedings. - 1613-0073. ; 2476, s. 138-153
-
Konferensbidrag (refereegranskat)abstract
- Background: Continuous Integration (CI) is an agile software development practice that involves producing several clean builds of the software per day. The creation of these builds involve running excessive executions of automated tests, which is hampered by high hardware cost and reduced development velocity. Goal: The goal of our research is to develop a method that reduces the number of executed test cases at each CI cycle.Method: We adopt a design research approach with an infrastructure provider company to develop a method that exploits Ma-chine Learning (ML) to predict test case verdicts for committed sourcecode. We train five different ML models on two data sets and evaluate their performance using two simple retrieval measures: precision and recall. Results: While the results from training the ML models on the first data-set of test executions revealed low performance, the curated data-set for training showed an improvement on performance with respect to precision and recall. Conclusion: Our results indicate that the method is applicable when training the ML model on churns of small sizes
|
|
| 3. |
- Al Sabbagh, Khaled, 1987, et al.
(författare)
-
Selective Regression Testing based on Big Data: Comparing Feature Extraction Techniques
- 2020
-
Ingår i: IEEE Software. - 1937-4194 .- 0740-7459. ; , s. 322-329
-
Konferensbidrag (refereegranskat)abstract
- Regression testing is a necessary activity in continuous integration (CI) since it provides confidence that modified parts of the system are correct at each integration cycle. CI provides large volumes of data which can be used to support regression testing activities. By using machine learning, patterns about faulty changes in the modified program can be induced, allowing test orchestrators to make inferences about test cases that need to be executed at each CI cycle. However, one challenge in using learning models lies in finding a suitable way for characterizing source code changes and preserving important information. In this paper, we empirically evaluate the effect of three feature extraction algorithms on the performance of an existing ML-based selective regression testing technique. We designed and performed an experiment to empirically investigate the effect of Bag of Words (BoW), Word Embeddings (WE), and content-based feature extraction (CBF). We used stratified cross validation on the space of features generated by the three FE techniques and evaluated the performance of three machine learning models using the precision and recall metrics. The results from this experiment showed a significant difference between the models' precision and recall scores, suggesting that the BoW-fed model outperforms the other two models with respect to precision, whereas a CBF-fed model outperforms the rest with respect to recall.
|
|
| 4. |
- Antinyan, Vard, 1984, et al.
(författare)
-
Monitoring Evolution of Code Complexity and Magnitude of Changes
- 2014
-
Ingår i: Acta Cybernetica. - 0324-721X. ; 21:3, s. 367-382
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Complexity management has become a crucial activity in continuous software development. While the overall perceived complexity of a product grows rather insignificantly, the small units, such as functions and files, can have noticeable complexity growth with every increment of product features. This kind of evolution triggers risks of escalating fault-proneness and deteriorating maintainability. Goal: The goal of this research was to develop a measurement system which enables effective monitoring of complexity evolution. Method: An action research has been conducted in two large software development organiza-tions. We have measured three complexity and two change properties of code for two large industrial products. The complexity growth has been measured for five consecutive releases of products. Different patterns of growth have been identified and evaluated with software engi-neers in industry. Results: The results show that monitoring cyclomatic complexity evolution of functions and number of revisions of files focuses the attention of designers to potentially problematic files and functions for manual assessment and improvement. A measurement system was developed at Ericsson to support the monitoring process.
|
|
| 5. |
- Ochodek, Miroslaw, et al.
(författare)
-
LegacyPro: A DNA-inspired method for identifying process legacies in software development organizations
- 2020
-
Ingår i: IEEE Software. - 0740-7459 .- 1937-4194. ; 37:6, s. 76-85
-
Tidskriftsartikel (refereegranskat)abstract
- Changing a software development process is a tricky task—the bigger the change, the trickier it gets. Large companies have the inertia of processes, the change of process takes time, happens over multiple releases and at different pace in different parts of the organization. Unfortunately, there are no effective tools available that help us determine if an organization has really adopted a proclaimed process change, as well as to what extent it is making progress towards this desired state. This paper presents a novel and unique method for determining the factual adoption of new processes in software R&D organizations. We use a DNA-inspired analysis (motifs) to categorize parts and find similarities between projects using defect-inflow profiles. We applied the method to analyze projects from a large infrastructure provider and from open source and show quantification of the evolution of processes. IEEE
|
|
| 6. |
- Ochodek, Miroslaw, et al.
(författare)
-
Mining Task-Specific Lines of Code Counters
- 2023
-
Ingår i: IEEE Access. - 2169-3536. ; 11, s. 100218-100233
-
Tidskriftsartikel (refereegranskat)abstract
- Context: Lines of code (LOC) is a fundamental software code measure that is widely used as a proxy for software development effort or as a normalization factor in many other software-related measures (e.g., defect density). Unfortunately, the problem is that it is not clear which lines of code should be counted: all of them or some specific ones depending on the project context and task in mind? Objective: To design a generator of task-specific LOC measures and their counters mined directly from data that optimize the correlation between the LOC measures and variables they proxy for (e.g., code-review duration). Method: We use Design Science Research as our research methodology to build and validate a generator of task-specific LOC measures and their counters. The generated LOC counters have a form of binary decision trees inferred from historical data using Genetic Programming. The proposed tool was validated based on three tasks, i.e., mining LOC measures to proxy for code readability, number of assertions in unit tests, and code-review duration. Results: Task-specific LOC measures showed a "strong" to "very strong" negative correlation with code-readability score (Kendall's $\tau $ ranging from -0.83 to -0.76) compared to "weak" to "strong" negative correlation for the best among the standard LOC measures ( $\tau $ ranging from -0.36 to -0.13). For the problem of proxying for the number of assertions in unit tests, correlation coefficients were also higher for task-specific LOC measures by ca. 11% to 21% ( $\tau $ ranged from 0.31 to 0.34). Finally, task-specific LOC measures showed a stronger correlation with code-review duration than the best among the standard LOC measures ( $\tau $ = 0.31, 0.36, and 0.37 compared to 0.11, 0.08, 0.16, respectively). Conclusions: Our study shows that it is possible to mine task-specific LOC counters from historical datasets using Genetic Programming. Task-specific LOC measures obtained that way show stronger correlations with the variables they proxy for than the standard LOC measures.
|
|
| 7. |
- Ochodek, Miroslaw, et al.
(författare)
-
On Identifying Similarities in Git Commit Trends—A Comparison Between Clustering and SimSAX
- 2020
-
Ingår i: SWQD 2020: Software Quality: Quality Intelligence in Software and Systems Engineering. - Cham : Springer. - 1865-1348 .- 1865-1356. - 9783030355104
-
Konferensbidrag (refereegranskat)abstract
- Software products evolve increasingly fast as markets continuously demand new features and agility to customer’s need. This evolution of products triggers an evolution of software development practices in a different way. Compared to classical methods, where products were developed in projects, contemporary methods for continuous integration, delivery, and deployment develop products as part of continuous programs. In this context, software architects, designers, and quality engineers need to understand how the processes evolve over time since there is no natural start and stop of projects. For example, they need to know how similar two iterations of the same program or how similar two development programs are. In this paper, we compare three methods for calculating the degree of similarity between projects by comparing their Git commit series. We test three approaches—the DNA-motifs-inspired …
|
|
| 8. |
- Ochodek, Miroslaw, et al.
(författare)
-
Recognizing lines of code violating company-specific coding guidelines using machine learning A Method and Its Evaluation
- 2020
-
Ingår i: Empirical Software Engineering. - : Springer Science and Business Media LLC. - 1382-3256 .- 1573-7616. ; 25, s. 220-265
-
Tidskriftsartikel (refereegranskat)abstract
- Software developers in big and medium-size companies are working with millions of lines of code in their codebases. Assuring the quality of this code has shifted from simple defect management to proactive assurance of internal code quality. Although static code analysis and code reviews have been at the forefront of research and practice in this area, code reviews are still an effort-intensive and interpretation-prone activity. The aim of this research is to support code reviews by automatically recognizing company-specific code guidelines violations in large-scale, industrial source code. In our action research project, we constructed a machine-learning-based tool for code analysis where software developers and architects in big and medium-sized companies can use a few examples of source code lines violating code/design guidelines (up to 700 lines of code) to train decision-tree classifiers to find similar …
|
|
| 9. |
- Ochodek, Miroslaw, 1980, et al.
(författare)
-
Using Machine Learning to Design a Flexible LOC Counter
- 2017
-
Ingår i: Workshop on Machine Learning Techniques for Software Quality Evaluation. - : IEEE. - 9781509065974
-
Konferensbidrag (refereegranskat)abstract
- Abstract—Background: The results of counting the size of programs in terms of Lines-of-Code (LOC) depends on the rules used for counting (i.e. definition of which lines should be counted). In the majority of the measurement tools, the rules are statically coded in the tool and the users of the measurement tools do not know which lines were counted and which were not. Goal: The goal of our research is to investigate how to use machine learning to teach a measurement tool which lines should be counted and which should not. Our interest is to identify which parameters of the learning algorithm can be used to classify lines to be counted. Method: Our research is based on the design science research methodology where we construct a measurement tool based on machine learning and evaluate it based on open source programs. As a training set, we use industry professionals to classify which lines should be counted. Results: The results show that classifying the lines as to be counted or not has an average accuracy varying between 0.90 and 0.99 measured as Matthew’s Correlation Coefficient and between 95% and nearly 100% measured as the percentage of correctly classified lines. Conclusions: Based on the results we conclude that using machine learning algorithms as the core of modern measurement instruments has a large potential and should be explored further.
|
|
| 10. |
|
|
