Greenstone: a platform for semantic digital libraries

This chapter illustrates the impact on a well–known digital library system – Greenstone – when it is moved from fixed modules and simple metadata–based structures, to open semantic digital library modules. This change has profound effects on the tools available to end–users to retrieve relevant content from the library, and an equally significant impact on the digital library (DL) architecture. Most current DL systems contain protocols for internal communication that define information exchange solely in terms of searching, browsing, and document retrieval. These communications reflect traditional user interactions in the library. However, this regimented approach results in inflexible systems that are difficult to extend to support other retrieval techniques. Furthermore, simple field–based metadata limits the ability of the DL to connect or disambiguate key items of information, impeding the precision of retrieval. Greenstone, an open source digital library toolkit that has developed over the last 10 years [14], forms the basis for the work described here. The software comes in two flavours: Greenstone 2 and Greenstone 3. The former exemplifies the classic form of digital library, with the added twist that (through UNESCO involvement) it is capable of running on primitive computing platforms that are common in developing countries (e.g. Windows 3.1 using Netscape 4). The latter is a reimplementation that is backwards compatible with the earlier version but far more ambitious in its goals. Particularly germane to the present work is the fact that it adopts an open protocol that works in tandem with a dynamic, componentised architecture [2]. In this article we describe the semantic aspects of the Greenstone 3 design and compare it with the earlier version, as a representative example of the archetypal approach. To illustrate key design elements, we draw upon three examples—an alerting service, ontology-enhanced representation and retrieval, and interoperability with a tourist information system—where many of the required user tasks and system features cannot be achieved through traditional digital library capabilities alone. These build upon the Greenstone 3 infrastructure. The details are developed in two steps: