In this paper we present a new approach for the development of hypertext information systems that avoids a number of traditional hypermedia authoring problems, and facilitates the maintenance of links and the reuse of information resources. The model is set-based, and integrates features from object-oriented software design models. This helps authors to make a comprehensive (and comprehensible) design before starting an implementation. We also briefly describe an authoring system that implements this model, and which uses an existing relational database system to store documents, links and meta-information.