The E-R Model Versus the DM Model
While a table in a normalized E-R-derived database represents an entity and its relevant atomic descriptors, tables in a DM-derived database are said to represent dimensions of the business rules of the enterprise. The meaning of business rule used here is somewhat different from that used by writers in the business rules community, where the term applies more directly to the declarative domain, range, uniqueness, referential, and other constraints you can specify in the database.
While advocates of implementing a normalized physical schema emphasize the flexibility of the model for answering previously undefined questions (see, for example, Inmon, 2000), DM advocates emphasize its usability because the tables in a DM database are configured in a structure more akin to their business use.
The E-R model for an enterprise is always more complex than a DM model for the same enterprise. While the E-R model might have hundreds of individual relations, the comparable DM model typically has dozens of star join schemas. The dimension tables of the typical DM‑derived database are often shared to some extent among the various fact tables in the database.