Are there arguments against ER models and their extensions

ER modeling is not rich enough


ER modeling is too flat

Opponent: ER modeling leads to diagrams which are of high granularity. Schemata of high granularity are not surveyable at all. Thus you need better structuring within the schemata.

Proponent: For this reason, we developed the approach of ''design by units''. The schema is clustered into a set of sub-schemata. The clusters can be developed on their own. Clusters have a limited set of associations. Such schemata can be much better surveyed and are easy to extend.

Opponent: In this case you can use approaches developed by Allworth et. al, Feldman/Miller, Gandhi et. al, Gilbert, Martin, Moody, Simsion and Teorey as well?

Proponent: Design by units generalizes these methods. Thus, you can use these methods as well. Therefore ER modeling allows to handle a consistent set of abstraction layers at the same time.