Assessing the testing skills transfer of model-based testing on testing skill acquisition

Felix Cammaerts,Monique Snoeck
DOI: https://doi.org/10.1007/s10270-023-01141-1
2024-01-23
Software & Systems Modeling
Abstract:When creating a software model, it is necessary that it accurately captures the desired behaviour, while at the same time ensuring that any undesired behaviour is excluded. On the one hand, formal verification tools can be used to check the internal consistency of a software system, ensuring that the behaviour of one software component does not contradict another. On the other hand, software testing is essential to check the external validity of the model more comprehensively. Unfortunately, software testing is often overlooked in curricula, resulting in graduates with inadequate software testing skills for industry. Software testing tools such as TesCaV can be used to help teachers teach software testing topics in a non-intrusive and less time-consuming way. Previous research has shown that TesCaV is easy to use and that novice users produce better quality software tests when using TesCaV. However, it has remained unclear whether learners retain the skills they gain from using TesCaV even when the tool is not offered for help. In order to understand the positive effect of TesCaV on learners' software testing skills, this study conducted an experiment with 45 participants. The experiment used a pretest-treatment-posttest design. The results show that participants feel equally confident about the completeness of their test coverage, even though they identify more test cases. It is concluded that for course design, a capsule such as TesCaV can help students to understand the full complexity of software testing and help them to be more systematic in their approach.
computer science, software engineering
What problem does this paper attempt to address?