A pragmatic approach to bitemporal databases: conceptualisation, representation and visualisation

Abstract

In this paper, we describe a pragmatic approach to bitemporal data modelling, logical representation and visualisation. Our proposed approach, herein referred to as the Bitemporal Object, State and Event Modelling Approach (BOSEM A), enhances an event-oriented to temporal databases and uses enhanced RUBIS-Schema event concepts for representing bitemporal event types. In this approach, a typical bitemporal object type can be perceived as a composite of its bitemporal event types and bitemporal state types. We contend that time concepts and semantics of time-varying information in bitem­poral data models should be captured using valid time and transaction time attributes defined on bitemporal event types. At the logical representation level, our approach relies on the semantic knowledge embedded in temporal relationships between object types for deriving bitemporal historic states at runtime. We show how bitemporal database systems utilis­ing this approach can capture bitemporal event data and derive bitemporal historic state details using built-in knowledge rules. The knowledge rules represent the semantic assumptions associated with the stored bitemporal event data. In order to help users visualize bitemporal data, we describe operators on bitemporal databases that present different bitemporal database views to the user. The different user views can be graphically presented to the user in the form of screen objects. This work is based on our experience in the development of a prototype bitemporal database system called the Bitemporal Object Viewing System (BOVS)). In the BOVS, bitemporal database objects are graphically depicted as collections of active bitemporal screen objects representing bitemporal event and bitemporal historic state objects.