Mathias Loesch

mathias.loesch@uni-bielefeld.de

Bielefeld University Library

During the recent years, OAI service providers have become a well-established backbone of scholarly communication. However, in the meantime the development of the academic web did not stand still. Aggregating metadata and making them searchable via a unified interface is performed by many search engines and not unique anymore. Rather, a service provider could exploit the potential of Open Archives by supporting value-added services such as enrichment, refinement or data linking. Thus, when service providers become data providers themselves by offering the aggregated -- and ideally enhanced -- data via APIs, the development of further services will be fueled. In this way, the network of scholarly communication shall become richer and generate novel, maybe unforeseen applications of the data from local repository systems. Current activities in the context of BASE-SEARCH.NET focus on the re-design of the architecture for an OAI service provider to fulfill these novel requirements. Having been run by Bielefeld University Library since 2004, BASE-SEARCH.NET is a registered service provider that collects, normalizes, and indexes data using OAI-PMH and provides one of the world's most complete aggregations of OAI publication data. BASE-SEARCH.NET is providing data interfaces and index access for numerous other service providers. Moreover, we are developing (DFG funded) machine learning-based text categorization services to automatically apply subject indexing of OAI DC records for being able to provide access to subject-specific subsets of the aggregation. On the basis of these activities and experiences, we will present our ideas of a novel, open architecture for OAI service providers. Current activities in the context of BASE-SEARCH.NET focus on the re-design of the architecture for an OAI service provider to fulfill these novel requirements. Having been run by Bielefeld University Library since 2004, BASE-SEARCH.NET is a registered service provider that collects, normalizes, and indexes data using OAI-PMH and provides one of the world's most complete aggregations of OAI publication data. BASE-SEARCH.NET is providing data interfaces and index access for numerous other service providers. Moreover,we are developing (DFG funded) machine learning-based text categorization to automatically apply subject indexing of OAI DC records for being able to provide access to subject-specific subsets of the aggregation. On the basis of these activities and experiences, we will present our ideas of a novel, open architecture for OAI service providers.