From the beginning, the CDEFD modelling effort was torn between two contradicting approaches. The purist's approach to information modelling asks for an analysis of the problem domain entirely independent of the implementation, because techniques and theoretical approaches change rapidly. On the other hand, people in need of organizing their data here and now call for a function-oriented way of modelling, providing rapid solutions to imminent problems.
The model as here presented is a compromise between these approaches. On the one hand, it is the result of a comprehensive and time-consuming effort of data analysis and definition. For example, naming and delimitations of the major entities in the model alone went through more than a dozen profound version changes in the course of the project. However, the rules of modelling, particularly those defined by Chen (1976) for Entity Relation Modelling were somewhat bent to achieve results directly significant for implementation efforts: With the possible exception of subtypes, entities in the model directly represent relational tables, keys are defined, and n-to-m relationships are resolved in cases where they have attributes apart from the foreign keys. In some cases, even implementation shortcuts are incorporated into the model.
Names of data items and entities are italicized where they are explained in the text. Entity names are always cited capitalized. Entities may be duplicated in diagrams accentuating specific aspects of the model.