Entities and relations for historical relational databases

Research to extend models of data to handle the temporal dimension has been conducted mainly in the context of the relational data model. In the relational model, the primary object of database design, manipulation and retrieval is the relation, viewed extensionally as a finite set of tuples. This paper proposes three useful types of temporal entities: events, histories, and valid-time relations, and defines relations and operations which allow for patterns of ordering and duration information to be defined and extracted. These patterns can often be expressed as select conditions on temporal relations in an algebraic query language. This paper is both a survey of existing trends in temporal database research, and a refinement of the conceptual framework used to characterize temporal information.