Introduction

Hypertext systems have found various practical applications that can range from on-line documentation, information retrieval systems, to sophisticated learning environments. These applications fall into four general classes: browsing systems, problem exploration tools, macro-library systems, and general hypertext systems (Conklin, 1987).

Hypertext systems, compared to traditional information retrieval systems, provide users with an easy and flexible access to a large amount of information. Hypertext proponents claim that the most salient advantages of these systems are the modularity of information and non-linear access to information through linking (Mohages, 1992).

The primary method for navigating through a hyperdocument is by browsing. However, the limitations of the browsing paradigm were soon revealed when it deals with large hypertext systems. The well-known problems of disorientation and cognitive overload in hypertext systems have been frequently reported and discussed in hypertext literature (Chen & Stanney, 1999; Conklin, 1987; Halasz, 1988; Nielsen, 1996; William, Eveland & Dunwoody, 2001). Marchionini (1988) states that the problem of disorientation will likely to diminish as the user gain experience with the system and as the designers apply common-sense interface designs. Consequently, designers must not only consider how to structure knowledge from a system performance standpoint, but must also consider what views and corresponding navigational tools are provided to the users.

Providing visual tools, such as overview diagrams or maps, is usually considered an efficient way of helping users navigate through complex hypertext structures (Halasz, Moran, & Trigg, 1987; Yankelovich, Meyrowitz & Druncker, 1988). However, as noted by many hypertext authors (Bernstein, Garzotto, Paoloni & Schwab, 1991), designing good overview diagrams for complex structures has proved to be difficult. Overview diagrams may facilitate navigation well in small hypertext systems, but for large systems, overview diagrams might introduce navigational problems of their own (Nielsen, 1990).

Hypertext usage depends on the mental models users have for the system. These mental models, in turn, depend on the conceptual model used by designers to create the system. The system views and navigational tools will be assimilated into a mental model for a system if they are familiar (Marchionini & Shneiderman, 1988). Therefore, designers must know how users seek information in traditional print systems and in the existing electric systems if they are to produce effective interfaces for new systems.

Many readers of books make extensive use of the table of contents and indexes to navigate through a book. They may transfer this experience to using hypertext systems. Table of contents (TOC) shows how the content of a book is related to its structure and provides the terminology of the book grouped in the context of its use (Jacques, Nonnecke, Preece & McKerlie, 1993). Table of contents is a valuable tool for presenting not only the structure of a hypertext system (Simpson & McKnight, 1990), but the relationships between items within the documents as well. If they are present in hypertext systems, they can provide tremendous help to users.

The interface design for table of contents can vary in hypertext systems. The TOC may be truncated if viewing the response to a query, and may use a fisheye view (Furnas, 1986). An expandable TOC, for example, presents the structure of a hypertext system by employing a fisheye-view method. The expandable TOC, while containing only the highest hierarchical level of headings when a document first displays, allows users to expand each section to its next lowest level and open as many different parts of the table of contents as desired at the same time. Types of interface chosen for table of contents can affect user behaviors, navigation patterns, and ultimately user performance and satisfaction with hypertext systems.