A formal basis for dynamic schema integration

Abstract

We construct a framework for determining when two schemata can be meaningfully integrated. Intuitively, two schemata can be integrated if the rules of one of the schemata together with a set of integration assertions do not restrict the models of the other schema. We formalise this concept using the notion of conflictfreeness and shows how it can be used to ensure that the merging of two schemata results in an integrated schema with the same information capacity as the original ones. The problem of conflictfreeness is undecidable, and we outline how it can be addressed for finite domains and determine its complexity properties in this case. Our approach takes into account static as well as dynamic aspects of a schema, where the dynamics is modelled by means of the event concept. We introduce correspondence assertions for events that can be used to specify the equivalence of event combinations.

This work was partly performed within the project SDeLphi at Telia Research AB. It was also partly supported by the NUTEK project CISIS.