Indexing and retrieval in digital libraries : developing taxonomies for a repository of decision technologies

Abstract

DecisionNet is an online Internet-based repository of decision technologies. It links remote users with these technologies and provides a directory service to enable search and selection of suitable technologies. The ability to retrieve relevant objects through search mechanisms is basic to any repository's success and usability and depends on effective classification of the decision technologies. This thesis develops classification methods to enable indexing of the DecisionNet repository. Existing taxonomies for software and other online repositories are examined. Criteria and principles for a good taxonomy are established and systematically applied to develop DecisionNet taxonomies. A database design is developed to store the taxonomies and to classify the technologies in the repository. User interface issues for navigation of a hierarchical classification system are discussed. A user interface for remote World Wide Web users is developed. This user interface is designed for browsing the taxonomy structure and creating search parameters online. Recommendations for the implementation of a repository search mechanism are given.