Person: Hassan Collado, Samer
Universidad Complutense de Madrid
Faculty / Institute
Ingeniería del Software e Inteligencia Artificial
Lenguajes y Sistemas Informáticos
Now showing 1 - 1 of 1
PublicationHacia una mejora en la corrección automática parcial de actividades/prácticas de los estudiantesGarcía-Magariño García, Iván; Arroyo Gallardo, Javier; Bravo Agapito, Javier; Galvez Gutiérrez, Daniel; Gómez Sanz, Jorge J.; González de Miguel, Ana M.; Hassan Collado, Samer; Herrero Desvoyes, Hugo; Lacuesta Gilaberte, Raquel; López Fernández, Marta; Nevzorov Oussenko, Aitor; Palero San Román, Inés; Pavón Mestras, Juan; Pita Andreu, Isabel; San Martín Doblado, Laura; Sánchez Hernández, Jaime; Segura Díaz, Clara María; Trillo Carreras, Juan; Vukotic de la Puente, Lucas; Yllana Santiago, DanielThe evaluation of student learning is usually done through activities and exercises where students apply their acquired knowledge in the resolution of problems close to real life, but delimited by the teacher. In these activities and exercises the teacher can evaluate conceptual knowledge and procedural knowledge. In the application area of this project, it is common that in the programming exercises and activities (practices) both types of knowledge are evaluated, since, on the one hand, the application of algorithms (procedural knowledge) and what programming elements and structures are being used (conceptual knowledge) are evaluated. The teacher has the complex task of evaluating this knowledge by making an individualized correction of each practice. While individualized correction of each practice is usually unbeatable in terms of the quality of the correction, on the other hand, automating the evaluation of procedural knowledge provides certain advantages. Routine parts can be evaluated automatically and thus save valuable time for the teacher's correction of other aspects. One of the advantages of automatic correction is that it is immediately available to the students and allows guiding them in essential aspects. In this teaching innovation project, a new online judge called UnitJudge has been developed. This new judge allows automatically evaluating programming practices in a consistent way even when these are long. In the ability to evaluate long practices, it outperforms other existing judges, such as DomJudge, which are more appropriate for short exercises since they are based on global inputs and outputs without allowing students to know which part of the code is failing. The newly developed judge allows to test the different parts based on unit tests. It is implemented for both C++ and Java practices. The new UnitJudge was used in several groups of the Programming Fundamentals subject. Taking into account the averages of 29 students' responses to the validated USE (Usefulness, Satisfaction and Ease of Use) scale, it was concluded that UnitJudge was easy to learn to use (mean of 5.99 out of 7), useful for the students (mean of 5.62 out of 7), and satisfactory for them (mean of 5.12 out of 7). The results of this teaching innovation project have been presented in two papers at the international conference "The 10th International and the 16th National Conference on e-Learning and e-Teaching" (ICELET 2023), respectively about (1) the presentation of UnitJudge and the experiment with the students of Fundamentals of Programming, and (2) intrusion detection from the viewpoint of cybersecurity in online judges exemplified with UnitJudge.