Working Paper Alfred P. Sloan School of Management a Metadata Approach to Resolving Semantic Conflicts a Metadata Approach to Resolving Semantic Conflicts a Metadata Approach to Resolving Semantic Conflicts

Semantic reconciliation is an important step in determining logical connectivity between a data source (databcise) and a data receiver (application). Semantic reconciliation is used to determine if the semantics of the data provided by the source is meaningful to the receiver. In this paper we describe a rule-bzised approach to semantic specification and demonstrate how this specification can be used to establish semantic agreement between a source and receiver. We describe query processing techniques that use these specifications along with conversion routines and query modification to guarantee correct data semantics. In addition, this work examines the effect of changing data semantics. These changes may occur at the source of the data or they may be changes in the specifications of the data semantics for the application. Methods are described for detecting these changes and for determining if the database can continue to supply meaningful data to the application. Though described in terms of the source-receiver model, these techniques can also be used for semantic reconciliation aind schema integration for multidatabase systems. Keywords[data dictionaries, metadata, query modification, schema integration, semantic conflicts]