LEARNING RESOURCE MANAGEMENT THROUGH SEMANTIC ANNOTATION FEATURES IN POPULAR AUTHORING SOFTWARE
Students are nowadays confronted with a disproportionately growing set of learning material. Finding, among the large number of resources provided by teachers in a LMS, the most appropriate ones is a non-trivial task for the learner. Extending the metadata of documents through the use of tags is a possible way of easing the filtering process. While automatic annotations may seem to be a pragmatic choice, the results may not always be suitable for every domain, due to the inherent ambiguity of natural languages. In addition, an LMS is not necessarily aware of the domain concepts that a learning resource covers, which could be critical to the filtering process. Also, current standards often apply metadata at a coarse-grained level. Tagging requires adequate support to not further increase both the authoring effort and the polysemy of the appended tags.
While most previous attempts realize a closed environment with limited to no interoperability, we propose an open ecosystem of plugins for popular authoring software practicioners use in their daily duties, to enable ontology-based tagging. We developed plugins for Microsoft Word and PowerPoint, as well as Adobe Acrobat. These plugins enable teachers to annotate learning resources based on tags from different existing sources of domain concepts. This will allow to distinguishably mark these documents as relevant for the set of appended concepts and thereby help the learner with quickly finding appropriate material. Tags are contained within the files and can be applied at a fine-grained level. In Word, tags can be applied to selections within a document, whereas in PowerPoint, teachers can add tags to individual slides. Learners are also provided with a list of tagged areas to quickly access relevant parts. Our Acrobat plugin allows to add tags to individual pages in a PDF document. The advantage over regular bookmarks is the use of common domain concepts. Autocompletion keeps the authoring effort low and fosters the convergence of concepts. Furthermore, the use of ontology concepts allows, depending on the domain, to automatically link to external resources providing further information.
Once documents have been annotated, an LMS can benefit from these tags. We provide Moodle plugin which, upon file upload, extracts available tags from the aforementioned document types and indexes them. Tags on resources and activities is an only recently introduced Moodle feature, together with visualization and filtering mechanisms. As tags are contained in the learning material and automatically extracted by an LMS, the resources become reusable in different instances or types of LMS without the need to manually reindex the tags in every LMS. Material already containing tags but not annotated using the previous plugins is also considered for automatic extraction, such as Idea Spaces from Zengobi Curio or Web 2.0 resources such as Stack Overflow Documentation.
Our approach supports the teacher in the authoring process without limiting her to a single authoring tool or LMS. At the same time, it enables the learner to quickly and effectively find appropriate resources among the vast set of material and, within these documents, find the information they need at a given point in time at a fine-grained level. We show the utility of the integration of these different tools through examples from different study domains to demonstrate the transferability of our approach.