Applying Functional Languages in Knowledge-Based Information Integration Systems

Abstract

Knowledge-based information integration systems exploit rich descriptions of domain knowledge to support query formulation, source reconciliation or query optimisation. A characteristic shared by most such systems is that at some stage during query compilation, a query is translated from an expression over the knowledge base into one phrased in terms of the data models of the (possibly wrapped) sources that are being integrated. Individual proposals differ (i) in the knowledge model used; (ii) in the source model and language used; (iii) in the nature of the rewriting required from the knowledge model to the source model; and (iv) in the role of the source model and language. This chapter discusses some of the alternative options and describes in some detail a knowledge-based query processor, in which: (i) the knowledge model is an expressive Description Logic; (ii) the source model is an object model and the source language is the monoid comprehension calculus; (iii) the rewriting involved is query expansion, in a global-as-view approach to schema integration; and (iv) the source language is used for query normalisation and semantic optimisation, but not for direct evaluation. The approach thus illustrates one role that a language with functional underpinnings can play in a knowledge-based integration system, and provides a context for the comparison of alternative functional source languages and roles for those languages.