RETRIEVAL HIERARCHIES IN HYPERTEXT

We have developed a collaborative, reuse hypertext system that has nov el browsing and retrieval characteristics. The system, called Many Usi ng and Creating Hypertext (MUCH), has been implemented on a network of UNIX workstations and used extensively in our group. This paper prese nts the model underlying the use of the MUCH system with respect to or ganizing, retrieving, and reorganizing information. In order to reuse information successfully, one must first organize it, then retrieve it , and finally reorganize it. The storage layer of hypermedia is logica lly based on nodes and links, and the MUCH system assumes that the nam es for these nodes and links represent a kind of semantic net. Documen ts, thesauri, and discussions may all be connected in this semantic ne t. The various functionalities of the system then exploit the knowledg e in this semantic net. Traversals of the net with various filters are the basis for the views that users get of the semantic net. The stand ard perspective is of a fold-unfold outline that represents a fisheye view of the semantic net. The reorganization of information also depen ds on a selection of nodes and their presentation within a connected s ubnet.