MAINTAINING DATA-DRIVEN RULES IN DATABASES

A new model with invariant-based language effectively handles data-dri ven rules in databases and uses the rules' inherent semantic propertie s and supporting mechanisms to meet high-level language requirements. It is an extension of the basic PARDES model developed by Opher Etzion in 1990 to support derivations and integrity constraints in databases . The model's invariant-based language, unlike other programming langu ages, can follow data-driven rules' semantic properties. Such rules ar e activated by modifications of data items in a database, and they pla y an important role in many applications that maintain complex relatio nships between data items or interdependencies between parts of the da tabase. Applications include expert systems, real-time databases, simu lations, and decision-support systems. The authors present requirement s for choosing an adequate programming style that uses data-driven rul es. These requirements are based on software-engineering criteria such as compatibility with a high-level language and verifiability of the rule language. The authors show that contemporary database programming styles fail to meet these requirements, and they present the invarian t-based language as a viable solution.