Assuming that private industry will take the lead in mainstream product development and short-term research, the steering committee presents here recommendations for research that it believes federal research agencies should encourage.¹ Federal initiatives that emphasize long-term goals beyond the horizon of most commercial efforts and that may thus entail added risk have the potential to move the whole information technology enterprise into new modes of thinking and to stimulate discovery of new technologies for the coming century. Of course, work should continue in current areas that have demonstrated promise, but the emphasis here is on opening up new opportunities. Five of the areas elaborated on under recommendations 2 and 3 in this chapter are designated for highest-priority attention because of their potential for contributing to the development of effective every-citizen interfaces.

Recommendation 1: Break away from 1960s technologies and paradigms. Major attempts should be made to find new paradigms for human-machine interaction that employ new modes and media for input and output and that involve new conceptualizations of application interfaces.

Needed in the technical community is a period analogous to the 1960s when a variety of paradigms were tried using emerging technologies of that time. The view then was of a single user interacting with a single terminal using a mouse, menus, and windows to open documents within a local application. The new view should take into account the tasks and technologies of the present and should enable a variety of interactions between humans and machines for communication, information retrieval, and performance of tasks in any of a variety of environments. Moreover, new approaches should enable users to immerse themselves in computer-mediated interactions and should include the option of involving many humans and machines in collaborative activities. New paradigms should emphasize the new role of information technology in society as a mediator among individuals, groups of individuals, and networked machines.

The steering committee's first recommendation is a call not to replace visual interfaces but rather to infuse them with new power and capability through document-centric design, speech, gesture, agents, position-aware and pressure-sensitive input devices, touch screens, and other emerging technologies and techniques. It also emphasizes bringing to fruition equally important new interface strategies, such as speech and voice response, that will carry the power of computing to environments and populations not served today. Today's interfaces often require too much of a user's vision and motor control in situations, such as driving, that present environmental distractions, or they assume physical or other abilities that many potential users may lack.

Coordinated research across several disciplines is necessary to develop new technologies and paradigms that address the psychological, organizational, and societal characteristics of every citizen. This interdisciplinary research should include the testing and evaluation of new interface technologies and paradigms in laboratory or field experiments or other empirical studies involving people who are representative of the citizenry.

The research agenda should acknowledge that the human-machine interface is more than screen deep and should consider every aspect of a person's experience in using computing and communications. People should be able to concentrate on the tasks or purposes for which they are using applications and should experience the interface as an aid rather than an obstacle to achieving success. People should experience a human-problem domain interaction rather than a human-machine interaction.

Several technological and design constraints should be considered in developing a research agenda:

• Architectures are needed for interfaces that have wide spectrum and are easily learnable. Such systems should have simple and semantically obvious commands so that novices can use them immediately. They also should have many levels of increasingly sophisticated capabilities that can be learned incrementally, so that people can move gracefully into more advanced and efficient uses. Such systems could be used in very sophisticated ways by experienced users.

• Systems should support modality and medium independence. Research should be fostered to enable anytime, anywhere, anyone interfaces that enable people to interact with systems using whatever modalities and media are available and convenient at a given time (consistent with the functions being performed). Such interfaces would support the goals of (1) ubiquitous and nomadic access so that a user can communicate on the road, down the hall from the office, calling in via telephone, and so forth; (2) equipment and communications system independence—use of low versus high bandwidth, one medium (e.g., audio, video, text) versus another, and so on; and (3) user ability independence, with special concern for people with disabilities and for the changes in abilities that typically accompany aging.

• Human-machine interfaces should support group (multiperson and multimachine) activities that are work oriented, social, or conducted for other purposes; the groups could be formal, established, or short term and ad hoc. In particular, interfaces should support communities of practice in which many individuals can participate, each contributing incrementally. To build interfaces that provide such support requires further development of theories of dialogue, theories of group behavior, and theories of joint planning and problem solving.

• Interfaces to the national information infrastructure should treat the two directions of communications—to and from individual citizens—more evenhandedly. Historically, with respect to elements of the NII such as broadcast television, citizens have been passive consumers of information. The evolving NII opens prospects for systems and interfaces that provide more flexibility in who can send and receive information, from what locations, and in what manner, as well as more flexibility and ease for people to move between communications- and information-centric activities.

• Information resources should have more attractive means of entry than those available so far. Two possibilities that could merit further research include (1) the concept of hyper-television that enables a viewer to pursue a presented object or event into cyberspace or (2) electronic encyclopedias with ubiquitous pointers into electronic libraries and other sources.

• Methods are needed that enable citizens to achieve the security and privacy they desire. Security-related features can be inconvenient to use; better interfaces could lower the barriers that have deterred their use historically.

Recommendation 2: Invest in the research required to provide the component subsystems needed for every-citizen interfaces. Research is needed that is aimed at both making technological advances and gaining understanding of the human and organizational capabilities these advances would support.

• Determine the needs of citizens: highest priority. Apply available sociological, psychological, and human-computer interface methodologies to try to understand the problems and the needs basic to effective human-machine interaction. Undertake studies to find the kinds of functionality and interfaces that will be most important. Such studies may be empirical or historical to determine what was successful in the past. For example, the usefulness and success of existing public information access projects, such as the Library of Congress system, could be examined. Proposed new technologies could be simulated and measurements made of their levels of success in real-life situations. Information so gathered should then guide decisions on what the most important technical areas are for research emphasis.

• Input technologies—Explore the promise of speech recognition and associated natural language processing: highest priority. Do the work necessary to open up speech as a viable input for as many new uses as possible. The steering committee is impressed that the range of potential applications for spoken input is tremendous, especially for hands-busy, eyes-busy situations; telephone applications; and differently abled persons. This need, coupled with the rate of progress in speech recognition, points to the importance of continued emphasis on this line of research. An important area that needs more attention is the construction of prototype speech interactive systems and their measurement and refinement in actual use.

• Improve understanding of computer vision, gesture sensing, and multimodal languages for user input. Computer vision can be used to gather data a user may wish to transfer to the network and to keep the system updated regarding the user's presence and responsiveness. Gesture recognition could involve the development of gesture languages and gesture support of multimodal languages.

• Measure the effectiveness of all of the above technologies when used by humans in problem-solving situations.

• Output technologies, including eyeglass displays; flexible, portable, and compact displays; high-resolution displays; virtual reality; haptic devices; mechanical actuators; voice and artificial sound; and multimodal generation of output. Develop display technologies to match human vision. Create audio output matched to the dynamic range of human hearing. Measure the effectiveness of these technologies in a systematic program to evaluate their relative strengths for human users.

• Ensure modality and medium independence: highest priority. Media independence and modality independence are goals that, to this time, have not been extensively researched and, in the new context, should be—for nomadic systems, for low-cost systems, and for people with disabilities. Develop mechanisms that enable translations between internal machine representations of information and various human representations (e.g., visual, audio, haptic). Research in this area will encourage the use of a common machine representation that can be flexibly translated into or from any available modes or media. Determine human multimedia and multimodal communications capabilities. Enumerate and prioritize human capabilities to modulate energy (as described in Chapter 3).

• Agent technologies. An important option for delivering services to users will involve agent technologies that interact with people to help determine their needs and then select domain-appropriate mechanisms to respond. The required technologies include traditional computer science mechanisms (such as those needed for database retrieval) and artificial intelligence capabilities (including representation of concepts and reasoning). An array of research topics needs to be addressed, including acquisition of user requests, user modeling, problem solving, and methodologies for summarizing and presenting internally stored data.

• Network access devices. A large variety of devices are needed, including mobile terminals, inexpensive minimalist systems, and full multimedia systems with virtual reality capabilities. The steering committee expects that industry will perform most of this work.

Recommendation 3: Encourage research on systems-level design and development of human-machine interfaces that support multiperson, multimachine groups as well as individuals.

• Develop theories and architectures for collaboration: highest priority. Develop theories of collaboration and problem solving. Develop architectures for networked people and machines that enable mutual awareness, easy communication across space and time, and individual and joint contributions to common goals. Provide ways to support community building and other social aspects of communication. Theories of collaboration have not been well developed historically, and the advent of networking makes this an important new priority for research.

• Human-centered design methodologies. Continue the study of human behavior in the use of technology for problem solving and the design of systems for improved productivity. Investigate the social effects of different interface choices, particularly the ways in which different presentation and communications choices affect people's interactions with media. To obtain feedback and to facilitate efforts at improvement, encourage social science research into how well the public is being served by such technology.

• Test proposed new designs: highest priority. Build experimental human-machine systems, for individual users or groups, using proposed technologies or simulations of them. Test, refine, and install them in applications environments, and measure their effectiveness. Industry, under the pressure of competition, has in recent years tended to minimize user testing in favor of quickly getting products to customers. Marketplace success has become the de facto test for usability by humans. Unfortunately, this approach does not lead to the kind of understanding that enables reasoned design of useful devices. Better understanding gained from testing and evaluation is needed to achieve breakthroughs to new paradigms, and to address the needs of differently abled individuals whose market buying power may be inferior to that of majority groups.

NOTE

1. In addition to supporting research, the government can encourage forward-looking approaches to accessibility for every citizen to the national information infrastructure by requiring adequate development processes and evaluation in the procurement and use of systems for public service facilities under government control.