A representational analysis on relationships in an ER model

We observe that a database is a type of representation system in that data in it represents something in the real world. Thus the notion of representation should be highly relevant to databases especially data modeling. In this paper, we focus on identifying the representation aspect of relationships of an ER model by means of notions of representation including information carrying and nomic constraint. In particular we review a known problem, i.e., connection traps, in ER modelling. Our main findings are: A transitive nomic constraint arises when the multiplicities of participating non-transitive relationships are *:1 or 1:1; Connection traps can be accounted for in terms of the nomic constraint or the lack of it in a path, namely, a fan trap is formed when a transitive structure contains no transitive nomic constraint, and a chasm trap is the result of the partial participation of an entity in the nomic constraint of a relationship. Furthermore, even when a path is free of connection traps, the transitive nomic constraint in the path has to project to (i.e., match) a constraint in the target domain (normally the real world) in order for the path to be able to represent it.