Ciência-IUL
Publicações
Descrição Detalhada da Publicação
Jinter: A hint generation system for Java exercises
ITiCSE 2023: Proceedings of the 2023 Conference on Innovation and Technology in Computer Science Education
Ano (publicação definitiva)
2023
Língua
Inglês
País
Estados Unidos da América
Mais Informação
Web of Science®
Scopus
Google Scholar
Abstract/Resumo
Programming novices often struggle when solving exercises, slowing down progress and causing a dependency on external aid such as a teacher, a more experienced person, or online resources. We present Jinter, a tool to generate hints to solve small exercises involving Java methods. The hints are produced taking into account the current state of an exercise and a backing model solution. The aid may refer to spotting errors or missing parts to achieve the desired outcome while taking into account behavioral equivalences of programming constructs (e.g., loop structures, forms of assignment, boolean expressions, etc). We evaluated the approach by surveying 8 programming instructors, finding that about two-thirds of the automated hints either match or are related to those given by instructors.
Agradecimentos/Acknowledgements
--
Palavras-chave
Introductory programming,Scaffolding,Hints,Feedback
Registos de financiamentos
Referência de financiamento | Entidade Financiadora |
---|---|
UIDB/04466/2020 | Fundação para a Ciência e a Tecnologia |
UIDP/04466/2020 | Fundação para a Ciência e a Tecnologia |
Projetos Relacionados
Esta publicação é um output do(s) seguinte(s) projeto(s):