We present a model for creating, managing, and viewing the contents of scientific hyperbooks. The model we propose is based on reusable information fragments and on terminological fragments that contain concept definitions. The definition of concepts, which plays an essential role in scientific writings, are represented in a formal language. These formal definitions can then be processed to automatically infer semantic links between fragments. In addition, concept definitions, which can be attached to “points of view”, can serve as a meta-information level to qualify the other fragments. Instead of having a fixed hyperdocument structure, as would be the case with HTML pages, the model includes an external view mechanism to generate a wide variety of hypertext views from the fragment repository. This enables the user to read (and write) the hyperbook in many different ways, according to different axis, needs, or tasks. A purely declarative language allows the definition of the views that make up the interface of the hyperbook. We also present the architecture of a hyperbook management system which is based on a database management system and a hypertext view generation system for databases.