Updating problem in semantic data models

Abstract

Semantic data models offer a rich set of abstraction mechanisms (e.g. generalization, aggregation and classification) to database designers to explicitly specify relationships amongst objects in a database system. Some of them also allow the definition of new relationships amongst objects. This results in a situation that the existences of and/or values of attributes of objects are highly depend on that of the other objects. For example, the deletion of one object may cause the deletion of other objects. Thus, updating must be carefully handled; otherwise, the database will be led into an inconsistent state. In this paper, the approaches taken by several semantic data models in handling updating are investigated; they include: functional data models, highly structured models, models focus on dynamic, and theory oriented models. Three distinct approaches in handling the updating problem are observed and discussed. To put semantic data model into a practical use, the execution of update operations must be under control.