Semantos : a semantically smart information query language

Abstract

Enterprise Information Integration (EII) is rapidly becoming one of the pillars of modern corporate information systems. Given the spread and diversity of information sources in an enterprise, it has become increasingly difficult for decision makers to have access to relevant and accurate information at the opportune time. It has therefore become critical to seamlessly integrate the diverse information stores found in an organization into a single coherent data source. This is the job of EII and one of the key components to making it work is harnessing the implied meaning or semantics hidden within data sources. Modern EII systems are capable of harnessing semantic information and ontologies to make integration across data stores possible. These systems do not, however, allow a consumer of the integration service to build queries with semantic meaning. This is due to the fact that most EII systems make use of XQuery, SQL, or both, as query languages, neither of which has the capability to build semantically rich queries. In this thesis Semantos (from the Greek word sema for “sign or token”) is proposed as a viable alternative: an information query language based in XML, which is capable of exploiting ontologies, enabling consumers to build semantically enriched queries. An exploration is made into the characteristics or requirements that Semantos needs to satisfy as a semantically smart information query language. From these requirements we design and develop a software implementation. The benefit of Semantos is that it possesses a query structure that allows automated processes to decompose and restructure the queries without human intervention. We demonstrate the applicability of Semantos using two realistic examples: a query enhancement- and a query translation service. Both expound the ability of a Semantos query to be manipulated by automated services to achieve Information Integration goals.