Executive Summary

INTERFACES FOR EVERYONE

Computing, communications technologies, and associated enterprises advanced enough in the early 1990s for the national information infrastructure (NII) to be accepted as public infrastructure. As a result, concern is growing about what will be required to enable most if not all of the public to use NII resources. The opportunity for broad public access and use reflects many factors, among them the technologies used directly by people as part of their interactions with the information and communications systems that make up the NII. This report outlines issues and directions for progress in developing interface technologies that will enable increasing numbers of people to use the NII effectively. Drawing from a late 1996 workshop hosted by the Computer Science and Telecommunications Board, the steering committee responsible for this project derived ideas for research in computing, communications, and social science to advance the underlying sciences and enable development of innovative, implementable concepts for interfaces that are more usable and capable than today's technologies and are accessible by as many people as possible.

The NII is dominated by computing and communications systems, and the human-machine or user interface represents the means by which people communicate with a particular system and the machines and people connected to it. In this report such technologies are referred to as every-citizen interfaces (ECIs), reflecting the project's mission to examine what might be required for every citizen to be able to use the resources available through the NII. ECIs are defined broadly as including input and output hardware and software as well as design and performance characteristics of applications—such as ease and speed of communication—that influence the overall experience of a person or group of people working in a system. The concern of the study is that, even though the usability of systems has improved substantially over many years, current interfaces still exclude many people from effective NII access. Most obvious are individuals with physical and other disabilities, but as articles in even the national and business press attest, people without such distinguishing characteristics, even expert users of NII systems, experience difficulties that constrain or even preclude their full use of NII resources.

The steering committee emphasizes that effective technological research on and development of ECIs must be grounded in a well-considered understanding of the needs and behavior of people. Achieving ECIs is thus an interdisciplinary endeavor involving computing-related science and engineering disciplines as well as social science disciplines. Progress toward developing improved ECIs will require basic research in theory, modeling, and conceptualization; experimental research involving building, evaluating, and testing of artifacts; and empirical social science research assessing segments of the population and how people actually work with different systems. In all cases, data, methodology, and tools are themselves targets for research or research support.

Certainly, however, the needed ECI-related research discussed in this report accounts for only part of the challenge of making NII resources broadly accessible. Policies aimed at promoting universal access to the NII must be developed that address economic factors, such as a person's ability to pay for communication and information services and access devices, as well as social and psychological factors, such as organizational, family, and peer group support, and personal preferences. Although the importance of such factors is clear, examination of them is beyond the scope of this report, which focuses primarily on issues related to computing, information, and communications technologies.




TECHNOLOGIES FOR HUMAN-MACHINE COMMUNICATION BY EVERY CITIZEN

At this time and for the foreseeable future, enlarging the set of options for human-machine communication, not replacing older technologies with new per se, is a broad goal for ECI research. Making a full range of options available involves continued improvements in mainstream interface technologies, such as graphical direct manipulation interfaces and typed and menu-selected command line interfaces, as well as research on modes that are currently not widely available. Recent advances in the performance and the commercial deployment of speech recognition and natural language processing technologies, for example, indicate their promise for future interfaces. In the last several years, progress in research and in the development of commercial products points to credible prospects for software agents as aids to ordinary citizens. Major progress in virtual reality technologies and three-dimensional user interfaces generally suggests their near-term availability. A variety of ideas are being generated for collaboration and communications technologies. Other options such as gesture recognition, pointing devices, and haptic devices are being proposed and in some cases developed and installed. Together, these technologies provide a rich set of opportunities to create new human-machine interface paradigms for the coming years.

Just as the NII is more than a single entity, so also will ECIs be diverse and varied. No single interface can be used by absolutely everyone, because people differ in ways relevant to the design of the technologies they use and because for a given person the activities and conditions of use for technologies vary: graphical user interfaces are problematic for blind people and also for people driving cars, for example. Experts who have concentrated on meeting the needs of users with specific limitations have discovered not only that it is possible to achieve adaptations of conventional interfaces to suit those users (giving rise to prototypes and commercial systems), but also that such adaptations often prove attractive and useful to many others. Such experiences underscore the value of medium and modality independence for future ECIs. Research and experience with real systems show that cross-disability access is compatible with diversity in the look and feel of an interface and that providing for it does not imply compromising capabilities that are useful to people without disabilities. In the language of this project, aiming for use by every citizen can enhance use by ordinary citizens. Even the seemingly ordinary are heterogeneous: the general population varies greatly in computer skills (e.g., from novice to expert); in the ability to speak, read, and write English; in personal cognitive styles (e.g., from linguistic/verbal to spatial/visual); and in personal propensity for using complex technological gadgets.

Other motivations for ensuring the versatility and adaptability of user interface technologies in the NII context include the desirability of achieving nomadicity, the ability of people to use the NII effectively regardless of their location, and the quality of available computing or communications equipment and services, which may vary depending on whether users are on the road, in the office, or en route between locations. Another more commercial motivation for emphasizing versatile and adaptable interfacing is the drive by relevant businesses to produce mass-market technologies. Although in the early 1990s popular discussions of the NII focused on the use of personal computers (PCs) and the Internet as an NII component, more recent experiences with commercialization and public acceptance are leading to a broader view in the late 1990s of an NII accessed by PCs, telephones, televisions, and a myriad of consumer electronics and other devices. Some of these devices have primary purposes other than computing and communications but incorporate embedded systems that allow for connection with networked infrastructure and therefore integration with other computing and communications systems. This diversification further broadens the range of people, activities, and environments that can be supported by the NII and thus represents one of the requirements for effective ECIs; it underscores the value of certain kinds of research, such as research and development to lower costs or foster compatibility. Taken together, the evolving set of motivations for facilitating human-machine communication gives rise to a range of ECI desiderata: flexibility, adaptability, ease of learning, compatibility, affordability, and so on.




SYNTHESIZING A RESEARCH AGENDA

An important starting point in building a research agenda is to ask what people's computing, communication, and information needs are with respect to the NII and how these needs can be met. This approach involves studying people doing ordinary tasks with and without technological aids and asking how new technologies might improve the process. It can also include gathering data from existing applications as input to guide new designs. Proposed systems can then be simulated or built in prototype for testing, refinement, and evaluation. Work of this kind can in turn guide decisions in technical areas concerning specific perceived needs and new research goals. A complementary approach is to study technical areas to discover fundamental mechanisms that can serve in providing support when they are needed. The pursuit of the two paths together can lead to an eventual synthesis of truly usable and important new aids for future communities.

Generally with this view in mind, the steering committee crystallized three recommendations that are summarized here and are presented in detail in Chapter 7. The first is that a major new effort be launched to seek new paradigms for human-machine interaction. The research community recognizes the success achieved by technologies developed two or more decades ago but also sees many indications that better alternatives are needed and are possible. Among the drawbacks of or problems with current technologies are that they are too finely tuned to the peculiarities of the technologically elite, too inflexible for the variety of applications and environments that the NII will offer, and too inaccessible to ordinary users or to individuals with disabilities. New, better interfaces are needed that utilize a variety of technologies that are now becoming available, that are usable by a broader cross section of people, that take advantage of the opportunities afforded by the NII, and that emphasize the new role of technology in society as a mediator among individuals, groups of individuals, and networked machines. Elaborating on this recommendation, the steering committee in Chapter 7 specifies a series of properties and characteristics that it thinks new and improved interfaces should have, including learnability, modality and medium independence, and a strong capability for supporting group activities.

The steering committee's second recommendation encourages investment in research on the component subsystems needed for ECIs and emphasizes the importance of studies of human and organizational behaviors and ways that technology can support them. It encourages research on a variety of potentially useful technologies, including input technologies (such as speech recognition, natural language processing, computer vision, gesture sensing, and multimodal input languages) and output technologies (such as flexible, portable, and compact displays, high-resolution displays, virtual reality, haptic devices, mechanical actuators, voice and artificial sound, and multimodal generation of output) that can help to maximize human-machine communication by more closely matching machine audio, visual, and mechanical capabilities to those of humans. This second recommendation also emphasizes the importance of developing modality and medium independence so that individual systems can be used by a variety of people in a variety of situations, and it recognizes the importance of agent technologies that can aid in interpreting and responding appropriately to users' needs.

The steering committee's third recommendation encourages research at the systems level that assembles the many subsystem components referred to in its second recommendation. It encourages the development of theories and architectures for collaboration and problem solving; emphasizes continued studies in human-centered design methodologies and social science research into how well the public is being served by new and proposed technologies; and underscores the importance of building experimental human-machine systems to test, refine, and measure the effectiveness of various proposed systems.

Because they have not been emphasized in other research or may have unusual payoffs in achieving improved ECIs, the steering committee chose to designate some areas for highest-priority consideration and accordingly emphasizes the following:

1. Undertake psychological, sociological, and historical studies to determine the needs of every citizen in the context of the NII and thus to provide guidance to technologists concerning what needs to be created and what will not work for real users.

2. Encourage additional research on speech recognition and the associated natural language processing so that speech can become a viable option for input in a variety of NII-related applications. Speech recognition and natural language processing are each important; speech as an output option calls for research related to speech synthesis. The steering committee was impressed regarding both the broad need for such capabilities and the recent progress in the field supporting the hypothesis that speech will soon become usable for at least some interface applications.

3. Develop technologies that enable modality and medium independence for as many applications as possible, in order to support the goals of nomadicity, compatibility of interfaces with a variety of hardware types, and usability by people with disabilities.

4. Develop theories and architectures that support collaboration among networked people. The new opportunities offered by the NII will come to fruition only if technologies are developed that enable collaboration.

5. Build experimental human-machine systems, for individual users and groups, using proposed technologies or simulations of them. Test them, refine them, install them in applications environments, and measure their effectiveness.

The steering committee emphasizes the importance of there being a strong experimental component in upcoming studies.