Motivation and Background

Working mostly in a teaching environment, the issue of knowledge engineering is imminent. Knowledge -- in its various forms -- has to be represented and later on presented to different audiences. Presentation formats like those offered by popular office packages suffer from a "write-only" nature. Developing and maintaining courses over time means that knowledge pertinent to a course has to be captured in a more manageable form. Once it is encoded semantically, it is available for transformation and presentation in many different forms, a feature which is quite helpful when courses have to be taken over by other people or when one and the same content is to be used in different contexts. Using one shared content backbone avoids maintaining variant copies. Also, such highly-structured representations allow a much more disciplined approach to link to external resources like online articles or books. Links can be added in a much more fine-grained manner allowing the user to access details without having to resort to search engines. But ultimately it is the promise of the enormous flexibility for adapting to changes, especially in Internet related affairs which motivated us to bypass the usual we teach, but we don't do mindset and develop - in several iterations - a Topic Map server. It was supposed to hold the shallow knowledge (encyclopedic content: topics and relationships between them) and should allow users - among them our students - to view and navigate through this knowledge base.

This itself involved three different activities. Firstly, much conceptual work had to be done to design formal languages for a long-living knowledge repository which could host a significant amount of topic maps. Then all the relevant content, namely Internet protocols, XML related technologies, Internet security and administration issues and also the field of semantic web had to be put in topic map form. And thirdly, of course, the actual development of the software.

In the following, we present a brief introduction to knowledge engineering with Topic Maps. This provides the foundation for the next section which covers the server's functionality and architecture as it has evolved over time. We then highlight some of the technologies we used, showing neuralgic parts in the server configuration. Next we list some of the experiences before an outlook finally presents some of our future plans.