Bulletin of the Technical Committee on Data Engineering

Abstract

Heterogenous databases, with their syntactic and semantic differences, are an information legacy that will prob- ably be with us always. Despite our efforts to define standards for information interchange, e.g., the current efforts with XML, bringing information together so that it can be sensibly queried remains a difficult and ex- pensive task, usually requiring human assistence. This is the “data cleaning” problem. Data warehouses usually face this problem every day, in their role as repository for information derived from multiple sources within and across enterprises. Data