Reduce data model clutter – now!

Bert Hooyman

Logical data models quickly become densely populated for all but the simplest data domains. The purpose of model diagrams is to communicate, but an urge to create complete and correct models is counter-productive to this. In this talk, we’ll talk about the various parts of UML that allow us to simplify a data model diagram without loss of completeness.

More specifically, we’ll discuss refactoring techniques like applying the composition pattern as an alternative to using composite associations; the use of array-valued properties and the elimination of base classes form a model diagram.


Current job: chief architect for a global IT services provider. In earlier position: four years as enterprise information architect for Philips Corporate IT and more recently for Philips Enterprise Information Management, with overall responsibility for the enterprise information strategy, information architecture and resulting design artifacts.

Comments are closed.