Lost in hyperspace: How can designers help?

Abstract

Recent advances in technology and the increased availability of microcomputers make it likely that many documents will soon be distributed electronically. One approach to the electronic presentation of documents is to imitate those in the paper medium, hut there are a number of arguments against this approach. An alternative is to use a method which takes advantage of the facilities that the new medium can provide, such as hypertext.

It has been proposed that people form mental maps of documents that they read, which are used for navigating through the document. However, the increased flexibility of access which characterises hypertext makes the formation of such maps more difficult, and readers may become lost within the document.

The series of studies reported here investigated the influence of a number of variables on subjects' ability to (i) relocate information from within a hierarchical hypertext document, and (ii) construct a map of its structure. An additional measure of interest was the relationship between performance in the two tasks.

It was found that showing subjects a textual representation of a HyperCard document's structure significantly improved performance, but this was further enhanced when a graphical representation was provided. The addition of a ‘footprint’ to the representation was of benefit, helping subjects to answer the question “Where am I?” in relation to the document as a whole. A non-interactive representation of the structure was useful, but an interactive one was of greater benefit. Providing subjects with a record of the cards which they had accessed improved performance in the relocation task, and showing the order in which these had been accessed further enhanced performance. Efficiency of information relocation was positively correlated with the ability to construct a map of the document's structure.

The implications of these findings for the design of hypertext systems are discussed, together with suggestions for fut lire research.