Conceptual schemas are described using some preferred data model theory. A taxonomy of potential changes can be derived from that data model theory, but it has no great significance for the maintenance and evolution of a given schema because it is based only on the constructs of the data model theory. This current state of the art in conceptual modelling can be improved by learning from actual business cases. This paper describes a number of semantic change patterns observed in operational conceptual schemes. Several semantic changes are not accounted for in current taxonomies. It shows that studying evolution in operational schemas opens up new and promising ways of improving data model theories and current design practices.
展开▼
