The Concept of Representation Capability of Databases and its Application in IS Development

The representation capability of an information system in general and a database in particular seems an important and yet elusive concept, which is concerned with, in our view, how a database ever becomes capable of representing real-world objects accurately or otherwise. To explore how to approach and then define this concept, we explore what is meant and required by the statement that a database connection (i.e., a connection between database constructs such as entities in an Entity-relationship (ER) diagram and relations in a relational schema that are made available by a database) refers to, represents and accurately represents a real-world relation respectively. This approach is proven to be insightful and effective. We also find a sufficient and necessary condition for a database connection to be able to accurately represent a real-world relation, which is that the information content of the database connection includes the real-world relation. All these make the concept of representation capability of a database approachable and definable. Furthermore, another different and yet related concept, namely the representation capacity of a database, can also be defined based on the representation capability of a database, which is ‘all the real-world relations that can be represented by the constructs that are made possible and available by the database’. Our theoretical work draws on semiotics, the semantic theory of information presented by Dretske and the information channel theory by Barwise and Seligman, and our practical work involves an information system’s development.