Summary

In this first attempt we successfully designed a topic map. As we have seen, we had to make some design decisions. Some of them concern the application domain: what data to cover and to which extent. Topic Map data usually does not have a natural boundary; anything directly or indirectly related to a central theme may fit into the map. The second concern is mainly how to encode particular data: whether occurrences or specially typed inline occurrences are sufficient or whether it is better (and cleaner) to create dedicated associations.

Once we identified a practical procedure, the next step could be to formalize the structure and the extent of such a topic map. This is a job to be done with an Topic Map constraint language which we will discuss in the next installment.