An OAI Compliant Federated Digital Library

To federate collections with varying degrees of metadata richness across continents poses a number of challenging issues such as: how to build a rich unified search interface when there is a lack of controlled vocabulary, and how to federate collections in different languages. The focus of this paper is on the approaches we have taken to address these issues in Arc, an OAI compliant federated digital library. The first Arc implementation had addressed the issue related to the lack of controlled vocabulary in an ad hoc way, and had not addressed altogether the cross-language issue. One approach to address lack of controlled vocabulary is by proposing a unified vocabulary and mapping individual library vocabulary to the unified vocabulary. This approach, particularly when applied to the subject metadata field, has two major limitations: (1) The standard subject classification (unified vocabulary for the subject field) may differ in terms of levels and semantics from that of an individual library leading to low precision searches. (2) Adding of a new collection to the federation leads to the complexity of updating mapping tables. Moreover, the new library may differ significantly in terms of the unified subject classification necessitating an update to the unified subject classification scheme. In this paper, we propose and implement a novel user-centric approach that addresses the lack of controlled vocabulary issue for the subject field, and does not suffer from above limitations. For the cross-language issue, the focus is more on how to integrate an existing cross-language technique based on query translation into Arc to support searches across English and German collections.