The ontological perspectives of the Semantic Web and the metadata harvesting protocol: Applications of metadata for improving web search

This paper compares the implied ontological frameworks of the Open Archives Initiative Protocol for Metadata Harvesting (OAI-PNIH) and the World Wide Web Consortium's Semantic Web. Both initiatives respond to problems with current search engine technology, particularly the search engine's equation of "web page" with "document," which is inadequate for efficient retrieval. The Semantic Web aims to mitigate these problems by introducing semantic mark up, which provides highly granular access to semantically meaningful segments within entire web documents. This approach is continuous with indexing principles of special libraries and online databases: deep, granular access that reaches beyond the document to access core semantic elements. The OAI-PMH works through the provision of metadata records supplied by the data provider to a metadata repository. This model, therefore, emphasizes the document as the central ontological entity, and therefore resembles the "document access" model of traditional subject heading schemes.