Review by: Dreier, Matthias (2004-07-27)

User interfaces are typically built with software toolkits. The availability of such tools has been crucial to the success of interface design techniques. Myers et al. provide an overview of past and present user interface software tools, both successful and unsuccessful ones. The article also provides an outlook on emerging user interfaces, e. g. PDAs, cell phones, and embedded computing devices.

The most successful software tools are window managers and toolkits. Virtually all graphical user interfaces offer a standard set of dialogs, menus and widgets. Another successful software concept is hypertext. Although hypertext was invented in the 1960s it has not been widely recognized until the success of the World Wide Web. The Web is one example of a technology that created a new kind of interface. Today, e-mail clients, online shops, library interfaces, and all sorts of configuration dialogs are web-based.

Approaches that did not succeed include user interface management systems (UIMS), formal language tools, and model-based techniques. These techniques for automatic generation of user interfaces mostly failed because they were too difficult to learn even for expert programmers, or because the resulting user interfaces were not as good as conventionally designed ones.

The authors predict new kinds of user interfaces to emerge from ubiquitous computing, 3-dimensional virtual reality applications, and recognition-based techniques. Ubiquitous computing denotes the concept of small, embedded computing devices (such as PDAs, cell phones, smart home and office devices) communicating via wireless connections. Those devices usually have no keyboard or mouse input and have only a small screen and a low-bandwidth connection. In contrast, people get used to larger screens and high-bandwidth internet connection on their desktop computer. The challenge of user interface design will be to deal with these different modalities.

Myers et al. illustrate how software tools influenced user interface techniques. Especially the lessons learned from unsuccessful approached are valuable. The authors even show some of their own approaches that did not succeed in the past. The description of future user interfaces shows very clearly that the widespread desktop metaphor and the tools used to design desktop applications cannot be applied to future devices. However, the vision of 3-dimensional interfaces and recognition-based input devices is rather optimistic. Several 3D user interface toolkits have been proposed since the mid-1990s. Speech-recognition has been predicted for decades to replace keyboard input. Both technologies failed to live up to their promises. Wisely enough, the authors do not give a precise date of when these new interfaces will emerge.