Adaptive hypermedia courses are difficult to debug, validate and maintain. Logfile analysis is partly to blame. We propose a graph-based approach to both real-time student monitoring and logfile analysis. Students are represented at their current locations in a dynamically created map of the course. Selected parts of student user models are visually exposed, and more detail is available on demand. Hierarchically clustered graphs, automatic layout and focus+context techniques are used to keep visual complexity at a manageable level. This component has been developed for an existing AH course system. However we believe that our approach can be readily extended to a wide selection of adaptive hypermedia course systems, filling in an important gap during course creation and maintenance.
展开▼