<- Virtual Exhibitions in Informatics

Synergistic database design with an extended Entity-Relationship model



Entity-Relationship (ER) models and extended Entity-Relationship (EER) models has limitations and problems.

The problems are:

  • Entity-Relationship (ER) models and extended Entity-Relationship (EER) models require designers to distinguish between attributes and entities. This can cause downstream redesign when attributes and entities are mismatched.
  • Design work can’t be completed in the ER model alone. Designers have to use two different types of abstraction. Before the transformation of the ER model to relations designer work with ER diagrams, but after the transformation they work with relation schemes. These relations may need normalization. Normalization is done through a combination of decomposition and synthesis techniques.

Embley and Ling propose an approach in their paper “Synergistic Database Design with an Extended Entity-Relationship Model” that solves these two problems. Their approach can be summarized as:

  1. Capture the real-world semantics in an improved EER model.
  2. Transform the EER model to a normalized EER model.
  3. Generate normalized relations.

Embley and Ling introduce an improved EER model called E²R (read “E-squared-R”). E²R is based on the ER model. It has many ideas such as generalization/specialization and lexical entity sets from semantic data models.

In the E²R model designers don’t have to distinguish between attributes and entities and it also supports the normalization at the model level.

So the major problems of ER models can be solved with their introduced model. Synergistic database design occurs as a designer interactively manipulates an E²R diagram until it has the properties that are desired. Design is done by transforming a given E²R into a normalized E²R model that is guaranteed to generate normalized relations.

More information

David W. Embley

Tok Wang Ling

8. ER 1989: Toronto, Canada

Reference

Embley, David W./Ling, Tok Wang: Synergistic Database Design with an Extended Entity-Relationship Model, ER 1989, pp. 111-128