Harvesting: Broadening the field of distributed information retrieval

Abstract

This chapter argues that in addition to federated search and gathering (as by Web crawlers), harvesting is an important approach to address the needs for distributed IR. We highlight the use of the Open Archives Initiative Protocol for Metadata Harvesting, illustrating its use in three projects: OAD, NDLTD, and CITIDEL. We explain how traditional services can be extended in a user-centered fashion, providing details of our new: ESSEX search engine, multischeming browsing, and quality-oriented filtering (using rules and SVMs). We conclude with an overview of work in progress on logging and component architectures, as well as a summary of our findings. © Springer-Verlag Berlin Heidelberg 2003.