Supporting Relationships in Digital Libraries

Preface This thesis is submitted to the Norwegian University of Science and Technology (NTNU) for the doctoral degree " doctor scientiarum ". The work reported has been carried out during the time period 1998–2003 at the First of all I want to thank my supervisor, Professor Ingeborg Torvik Sølvberg, for providing me with a fine combination of freedom and guidance in the shaping of this work. The discussions we have had and the feedback I have received throughout these years have been invaluable for this work. I also want to thank my colleagues at the Department of Computer and Information Science for the friendly and helpful environment that has enabled me to do this work. Particular thanks to BIBSYS for allowing me to use their database in the application that was developed as a part of this work. Finally, I want to thank a very patient and understanding family. Abstract The motivation for this work is based on two recognized research issues for digital libraries. One is the need for interlinked and semantically rich information spaces where relationship information is of particular importance. The other is the service-oriented architecture of digital libraries. The digital libraries of the future will consist of smaller and independent systems that each will provide different functionality and access to different contents. This work defines and explores a service for managing and using explicit relationships – the Digital Library Link Service. The service implements an instance-oriented approach to relationships that enables any kind of typed relationship to be created among the information objects of digital libraries. The service can be used to create consistent information spaces on top of digital library repositories and enables an associative organization and retrieval of information objects. This work shows that the use of a fine-grained relationship model implemented as distributed objects enables distribution of the relationship network while still being able to support constraints and maintain consistency. The cost of this, however, is a complexity that can reduce performance and scalability due to the call latency of network communication. A prototype is developed that utilizes caching in order to solve this. Tests conducted show that this technique significantly contributes to the scalability and efficiency. This is particularly important when the relationship information is distributed across different processes with high call-latency in between. The work further presents a prototype application for enhancing bibliographic catalogues with the rich set of relationship types …