Towards a Conversational Agent to Support the Software Testing Education
Leo Natan Paschoal, Lucas Fernandes Turci, Tayana Uchôa Conte, Simone R. S. Souza · 2019
The training of professionals in the field of software testing is increasing its relevance in the past few years and, therefore, efforts in appropriate methodologies for the learning-teaching process in this context have been proposed and appreciated. The emergence of pedagogical models, such as flipped classroom and team-based learning, which demand from the students a previous study of the theory before the lecture, creates a concern: how to support the before class learning? Because of the hybrid nature of these pedagogical models, which means they mix elements from traditional and distance education, it is possible that the support mechanisms used in distance learning platforms, such as conversational agents, can be applied for this matter. At the same time in which the academic work tries carefully to provide a proper software testing formation, there are also many contributions being established regarding the training and non-formal learning. Improvement and personal training courses about criteria, tools, and software testing good practices are being created by teaching institutes and offered in Massive Open Online Courses platforms (MOOCs). However, in this type of course, in the absence of a teacher, the student might be in a situation where there is nobody available to answer their questions about the topic. In this paper, we propose the use of conversational agents in solving the problems and challenges which encompass the learning through MOOCs and hybrid models. A conversational agent, called TOB-SST is proposed to support software testing education. A viability study was conducted to understand the quality of the given answers by TOB-SST and the possibility of it serving as a learning support tool. The results indicate that it is promising to employ a conversational agent to guide student study.