Modeling Data Quality and Context Through Extension of the ER Model

Capturing data quality and context semantics at the early stage of database design is a critical issue for both database researchers and practitioners. As in traditional database design, users' quality and context requirements should be represented at the conceptual level. This paper first generalizes the issues governing data quality and context semantics to be an Interattribute-Relationship problem, and then examines the feasibility of the Entity-Relationship (ER) model as a solution. Investigation on alternatives using the existing ER constructs reveals that it is necessary to extend the ER model. An extension called Attribute-Relationship (AR) is proposed to extend the view of having relationship at the attribute level. The attributes that involve in the relationship are termed strong attribute and weak attribute. In the AR extension, the dual roles of the strong and weak attributes and the embedded existence dependence between them are represented through an identifying relationship. Finally, a range of integrity constraints related to this extension are presented.