Every model has a purpose and the quality of a model ultimately measures its fitness relative to this purpose. In practice, models are created in a piecemeal fashion through the construction of many diagrams that structure a model into parts that together offer a coherent presentation of the content of the model. Each diagram also has a purpose – its role in the presentation of the model - and this determines what part of the model the diagram is intended to present. In this paper, we investigate what is involved in formally characterizing this intended content of diagrams as
and show how doing this helps to improve model quality and support automation in the modeling process. We illustrate the approach and its benefits with a case study from the telecommunications industry.