The hypertext model described relies on graphical presentation and positioning to indicate the relationships between items. The paper begins with an outline of three models of hypertext: document-centered, browser-based, and spatial. Document-centered hypertext is the traditional model, in which references are embedded in documents, and users move between documents by selecting appropriate references. Browser-based hypertext provides a map or graphical overview of the connections between items. Spatial hypertext is an extension of the ideas embodied in browser-based hypertext. It allows the relationships between items to be indicated by a combination of proximity, shape and color, and grouping into higher-level composite structures.

The authors have developed an experimental system, VIKI, to provide a testbed for the spatial hypertext paradigm. VIKI’s model consists of visual symbols--references that can be manipulated and grouped; objects--nodes with content; collections--user-defined subspaces that can contain arbitrary arrangements of other items (including nested collections), thus allowing a hierarchy of spaces; and composites--structures that contain particular visual patterns of two or more items or collections.

Selecting the symbol for an object provides access to the object. Selecting the symbol for a collection provides access to a scrollable window into a subspace containing nested items. Users may scroll, resize, or zoom the window to reveal more or less of its contents. Composites are similar to collections, but have an associated type that refers to the particular arrangement or pattern of the items. VIKI provides interesting features for recognizing repeated patterns and helping users to define and manipulate them. The overall model allows users to build up a sophisticated visual language to describe and manage large collections of objects.

This clear and easy-to-read paper compares VIKI’s model and facilities to previous hypertext systems and gives clear examples to illustrate VIKI’s features. I recommend it as an excellent introduction to the possibilities of spatial hypertext.