Planning for Information Gathering A Tutorial Survey

We discuss the problem of information gathering which is that of integrating mul tiple heterogeneous information sources so they appear as a single information source that a user can pose queries on We characterize information gathering as being made up of two subproblems query planning and plan execution Query planning is the process of searching through the space of rewrites of the original query to nd rewrites containing only queries on available information sources Query execution is the process of executing those queries and dealing with execution failure We relate this general view of information gathering with various information gathering systems including the Information Manifold SAGE Occam Razor and TSIMMIS