Speaker: Herbert Van de Sompel
Affiliation: Ghent University, Belgium
Time and Location: 4:15 PM, B17 Upson Hall
Title: SFX/OpenURL and the Open Archives Initiative: Achieving Interoperability in Digital
Libraries via Low-barrier Standards
The problem of interoperability in digital libraries is about building coherent services across distributed heterogeneous collections, running on a wide variety of technological infrastructures. Efforts towards achieving interoperability have had disappointing impact, presumably because of organizational difficulties. The specifications demand that the partners embrace complex standards. Few organizations follow them. In SFX/OpenURL and Open Archives, this consideration is the starting point for the creation of low-barrier interoperability specifications as a means to achieve maximal buy-in by digital libraries. With this regard, both efforts seem to be on a path towards success. In case of
SFX/OpenURL, substantial proof is already available that emphasizes the fact that great user-services can be created based on simple technical specifications for interoperability.
SFX/OpenURL is a proposal aimed at enabling the delivery of context-sensitive extended services. When linking from a citation to the article represented by the citation, one needs to take into account the fact that multiple instances of that article may exist. Delivery of the appropriate instance depends on the context of the user. But this problem is not limited to linking from citations to full-text. It is equally relevant for other types of so-called extended services that link a record from whichever scholarly information resource to related information. For instance, when an institution subscribes to an abstracting and indexing database run by an intermediate, its users should be able to link from records in that A&I database to corresponding records in the version of the citation databases to which the institution subscribes. Or, the link-to-holdings feature, pointing at the user's OPAC system, that is very commonly used for A&I databases, should also be available for citations in journal articles. All kinds of linkages between electronic scholarly information resources should take into account the context of the user. In general, this is not the case for actual linking solutions. A simple solution to these problems known under the name the SFX-framework and the OpenURL will be discussed. Also, it will be shown how application of the concepts of the SFX-framework in the DOI/CrossRef linking environment can make the latter context-sensitive.
The Open Archives Initiative started as a forum to discuss and solve matters of interoperability between e-print archives. It's Santa Fe Convention described a low-barrier protocol allowing selective harvesting of metadata from distributed e-print archives. Over the course of the past months, the protocol has attracted the interest of many different communities in the scholarly realm, looking for basic solutions to achieve a level of interoperability amongst their repositories. As a result of this, the Open Archives Initiative decided to revise the protocol in order to make it applicable in a broad context. The resulting Open Archives Metadata Harvesting protocol addresses harvesting of metadata from the so-called "deep web": information stored in databases, for which no static web pages are available. The protocol builds on the concepts of data providers -- the parties that run repositories -- and service providers -- the parties that build services for data harvested from data provider's repositories.