Modifying the entity relationship modelling notation: towards high quality relational databases from better notated ER models
2017-11-22T02:40:23Z (GMT) by
This working paper proposes several modifications to the original ER modelling notation. The effort will help create ER diagrams (ERD) which are improved notation-wise and are of less confused, and which will in turn help design high quality relational database(Codd, 1970) models. Based on the modified ER modelling notation, the present ER to relational model transformation algorithm (Batini, Ceri, & Navathe, 1992; Ramez Elmasri & Navathe, 2007; R Elmasri & Navathe, 2011) will also be modified in order to obtain high quality database models from ERDs.
This paper will also be used as a supplementary document for empirical evaluations of the that investigates ways of designing high quality databases from ERDs. By a high quality database we presume a database semantically clear, complete, and accurate with respect to its predecessor ERD and is easy to reverse back to the ERD without any intuition.
To achieve the goal, an entire approach will be provided that consists of following components:
1. A modified notation scheme for ER modelling
2. A new ER to relational forward transformation algorithm
3. A criterion for assessing the quality of relational database schema (RDS)
4. A RDS to ER re-engineering algorithm
Accordingly, this paper presents the first component: “A modified notation scheme for ER modelling”. The rest of the components will be published soon.
Transformation algorithms are inadequate(Pieris & Rajapakse, 2012), but it is not the only reason for a poor database designs to be resulted. Many methodologies and notation schemes are available for ER modelling(Batini et al., 1992; Ramez Elmasri & Navathe, 2007; R Elmasri & Navathe, 2011; Song, Evans, & Park, 1995; Teorey, Yang, & Fry, 1986). However, we take the view that most of them hardly can model a successfully transferable ER diagram to the relational model. To our understanding, the reason is the existence of a poor relationship among the concepts in the ER notation, the ER modelling method, the transformation algorithm, and the relational model. The first step to solve the issue is to provide a solid ER notation and modelling methodology that can create a transferable ERD as we expected.