C. Grévisse, J. Botev, S. Rothkugel

University of Luxembourg (LUXEMBOURG)
In project-based learning, the large amount of heterogeneous learning material, retrievable from different sources, requires a significant filtering effort from students as well as constant switching between the material and the actual project. Most of the time, there is also no dedicated link between the task at hand and supporting learning material, so students would need to be able to precisely identify the problem and search for related resources, which may result in Meno's paradox of inquiry.

We therefore propose to move from a material-centric to a student-centric approach by integrating learning material into an advanced project-based learning support platform. Not only teachers, but also students shall be able to add both traditional material and resources from online platforms. An underlying umbrella ontology enables a cross-curricular retrieval of matching learning material.

As a main use case, we are considering the domain of computer programming due to its inherent structural properties. Previous attempts are focused on content sequencing in separate tools or the integration of exercises inside an LMS. Our approach, however, supports direct retrieval of programming-related material from within an Integrated Development Environment (IDE) such as Eclipse and thus allows for ad-hoc task support. The learning material constitutes an open corpus, as new resources can be added at any time. This includes slides, exercise sheets, personal notes, but also articles from online platforms such as Stack Overflow.

The proposed umbrella ontology maps abstract programming concepts to elements in the Java and C languages as well as tags on Stack Overflow. The ontology is extensible with respect to other programming languages and user-specific concepts, which could be of particular interest in a multilingual setting. It also enables a cross-curricular integration of learning material reusable in different contexts. Resources from other courses will be transitively accessible through the mapping between annotated aspects in the material, the aspects present in a coding task and the related concepts in the ontology. Material on similar aspects in a different programming language or basic material from another course required to understand an advanced issue can be proposed. This task-independent mapping helps avoiding fragmented learning and particularly improve global learners' understanding. A fine-grained mapping between elements in code and the related material is also possible. In a divide-and-conquer approach, different levels of detail can be explained, from low-level syntax constructs to high-level API interconnections or programming patterns.

Our ecosystem is comprised of plugins for commonly used software, such as Microsoft Word and the Eclipse IDE. The Word plugin being available on Windows, macOS and iOS and both teachers and students being able to add material to the ecosystem, the authoring and annotation work is distributed. Once new material is added and its contained annotations are indexed, bidirectional links between the material and code are inherently supplied through the related concepts in the ontology. Finally, the Eclipse plugin shows the related material of a selected aspect in the code and enables the user to rate the recommended resources. Based on explicit ratings, implicit behavioral data and cross-domain preferences, future recommendations can be enabled.