Design science research applied to difficulties of teaching and learning initial programming

Learning and teaching to program is an arduous task. It requires a lot of commitment, dedication, and passion from everyone involved. Programming courses have high dropout and failure rates. Throughout time, several educational research works have been carried out to study the different learning processes and characteristics of students. With this work, we present and describe our vision and model of teaching and learning of initial programming to minimize the problems. We present a technological tool, called HTProgramming (Help To Programming), which complements the teaching and learning process. This allows students to practice a wide variety of activities with immediate feedback, directly related to content and themes for learning programming. It allows the teacher to follow the whole process and students' results. Using a machine-learning (neural network) predictive model of student failure, it will allow the teacher to anticipate possible student failure and act quickly. In this paper, we apply the Design Scientific Research Methodology to tackle teaching and learning difficulties to initial programming. We also include the results and evaluation of the application. Students consider the application an important tool for their learning process. The student failure prediction model presents very realistic values.