Browse Books

Reviewer: Tom Carey

This short book has two parts. The first 60 pages are a useful summary of online help systems, and the last 40 pages are an overview of user interface design concepts as applied to these systems. Despite its cursory treatment of the material, this book fills a useful niche in the human-computer interaction literature. Its major contribution is that it outlines the design options open to designers and implementors of help systems and indicates the other sources available to guide their decisions. The coverage of online help begins with a chapter that describes the decisions to be made during the design of a help system. The following chapter illustrates how existing systems provide help, and the final chapter in this part outlines empirical research results that are applicable to help systems. In principle, one might want to reorder the material so that form follows function: that is, first discuss when and why people need online help, then show how these needs can be met. In practice, the author's arrangement works well: the audience for this book probably needs to see the variety of design choices first, in order to be convinced that there is a lot worth learning about online help. The book does this job commendably. I was not satisfied with the author's coverage of the different functions required of a help system. The promised “what works and what doesn't” was probably the wrong target, but more of “what works when” would have been useful. A summary of Rasmussen's malfunction analysis [1], for example, would have helped to explain the different situations to which a help system must respond (and to provide a context within which to evaluate the different forms in the previous chapters). The user sees a help system in the context of an application. Many of the design decisions discussed in the early chapters could not be made in isolation from the application. For example, decisions on screen format, access methods, and extensibility should be integrated with the design decisions in the application system. Kearsley does not adequately address this relationship between the application system and the help system, nor does he question the appropriateness of our frequent practice of treating help as a separate add-on function. How is the design process (or this book) affected if we consider help as an integral requirement of every application, rather than as something to be designed separately__?__ The second part of the book gives an overview of user interface design. The risk here is that a little knowledge could become dangerous. I would like to have seen more disclaimers about the cursory nature of the discussion, as well as some more explicit recommendations about the additional knowledge the readers need before engaging in the design steps. For example, the author does not mention work to validate user satisfaction questionnaires [2] or to analyze effective methods of collecting verbal protocols [3]. Flaws in the graphs and the glossary detract from the book's utility. The captions of some graphs, such as Figure 4.2, are incomplete, so the diagrams make no sense without the accompanying text. In the case of Figure 5.3, a graph that depicts improvements in a proposed help system, I was not sure whether the author thought a designer should construct such a table, or whether it was meant only to illustrate the analysis that should be conducted informally. In the glossary of 34 terms, I found 8 of the definitions misleading, in some cases because the author implicitly restricts their scope to commonly available help systems. For example, “bullet proofing” is not finding all the errors in a program, but rather protecting the user and the program from each other's unanticipated actions. A menu is not just a “list of program options”—it could be a list of file names, for example. This book would be a useful introduction for someone unfamiliar with the literature on human-computer interaction, such as a technical writer who is now being asked to produce an online help system. The book's brevity, despite its disadvantages, may be an asset in attracting readers who do not yet recognize the scope of the knowledge required for systematic design. If this work encourages them to read further in this area, it will have played a valuable role.