Hypertext is a system that exhibits structure in documents, or document sets, by using links between related text elements or text portions. These links can be followed to construct particular text revisions and to display relationships inherent in the text. The paper under review is a well-written, easily readable work dealing with certain aspects of the hypertext environment. The paper briefly outlines the existing hypertext approaches, with emphasis on the authors’ own NEPTUNE system. A multiperson situation is then considered where several cooperating writers help in generating and revising a particular document. The problems that arise in preserving consistency and avoiding conflicts between multiple versions of a common document are considered in detail.

Because the paper is also conceived of as an overview of the hypertext environment, it might have been interesting if the authors had supplied, in addition to the technical implementation details, a critical assessment of the usefulness of conventional hypertext systems. Hypertext manipulations are sometimes advertised as methods for augmenting people’s knowledge by making it possible to jump from topic to topic, thereby supplying free associations between related ideas. Unfortunately, links cannot be followed until they are somehow supplied, and while it is easy to conceive of objective links--say between a given document and other documents referenced by the original item--it is not clear that this type of objective linking has much to do with idea associations. Subjective links that deal with text understanding and text content should be very hard to supply, whether manually or automatically, and they of course are most important if a hypertext system is to be really useful. I would have liked some discussion of these essential questions.