Research Motivated by Social Applications of Information Technology
The diffusion of the Internet, combined with advances in basic computing and communications technologies, is poised to fundamentally alter the nature of information technology (IT) research. As IT continues to move from the relatively simple realm of back-office transactions processing and personal productivity-enhancement tools into less specialized, mass-market contexts that support electronic commerce (e-commerce), delivery of government services, and personal interactions, the set of problems that motivates IT research is continuing to change. Many of the new applications are social applications that serve groups of people in shared activities. Simple social applications support the collaboration of geographically dispersed groups of people engaged in a shared task, such as designing a new product or writing a report. More sophisticated social applications support a range of business, economic, and societal functions, such as manufacturing processes or distance education. Social applications tend to integrate IT into larger sociotechnical systems that involve people, organizations, and other technologies and that derive their functionality from the complex interactions of IT with nontechnical system elements. Many of the social applications comprise large-scale systems of the kind described in Chapter 3, but social applications of IT pose a number of additional interesting research problems, the solutions to which will require more explicit collaboration among IT researchers, end users, and researchers in other disciplines. Progress along purely technical dimensions, such as processing power, communi-
cations speed, and data storage densities, will no longer suffice; a more holistic view is needed (Brown and Duguid, 2000).
This chapter examines the increasing integration of IT into larger, social applications and the shortcomings of today's technology relative to a complex set of expectations. The first two sections lay the groundwork for the analysis by identifying the characteristics of social applications and the many challenges they present. Underlying this discussion is the idea that, because IT is proliferating in social applications, research on social applications should be expanded in amount, scope, and depth and, furthermore, that this new research will require approaches that are somewhat different from those taken in much of the more narrowly technology-oriented research that is common today.
The third section discusses ways in which interdisciplinary research can play an important role in this arena and identifies some initial steps in this direction. Just as scientific computing has benefited from closer interaction between technologists and natural scientists, so can the more-social applications of IT deployed today benefit from collaboration between technologists and social scientists (including experts in law and business as well as psychology, sociology, anthropology, and economics). The fourth section examines mechanisms for pursuing technical and nontechnical research that could increase understanding of social applications of IT and thereby enhance capabilities to design, develop, deploy, and operate them. Building on the groundwork laid in the present chapter, Chapter 5 identifies specific steps that could be taken to stimulate more of this type of research.
The development of appropriate mechanisms for funding and conducting research on the sociotechnical dimensions of IT systems will be a significant challenge. This work can build on some important foundations, notably research on human-computer interactions and computer-supported cooperative work. These existing research efforts are inherently multidisciplinary in outlook because they are concerned with the ways in which people relate to systems. Experience to date in these areas illustrates both the promise of social applications and the practical problems involved. Multidisciplinary research is always problematic because of the difficulties inherent in bridging the gaps separating different communities of researchers. Compounding these problems is the need implied by the concept of social applications to engage not only established researchers in other disciplines but also end users of IT systems who understand the context in which IT systems operate and directly confront problems of implementation, ease of use, performance, and operation. Many end-user organizations have little or no history of conducting research, especially IT-related research. New mechanisms may therefore
be needed, and some ongoing activities are suggestive of the types of structures that may be effective.
The discussion proceeds at a high level because it is intended to bridge research themes each of which could justify detailed examination. Such examinations have been provided already by the Computer Science and Telecommunications Board (CSTB) in more focused assessments. This chapter points to that other work, which established an intellectual history developed through separate engagements with segments of the research community. That this chapter echoes and amplifies ideas raised previously is important enough to be acknowledged explicitly: research ideas and suggestions for how to make progress in the field either recur or linger largely unaddressed because the problems are difficult, because the recommendations are aimed at subsets of the research community despite broader relevance, and/or because there is a lack of readiness, whether due to insufficient insight and understanding of the needs or inadequate capabilities. The committee recognizes that it has focused on difficult problems but believes that the time is right to address them; its recommendations are aimed at promoting both understanding and capability.
SOCIAL APPLICATIONS OF INFORMATION TECHNOLOGY
Thanks in large part to networking, IT has become a factor in large organizational constructs, whether whole enterprises, groups of enterprises that interact in commerce, or the overarching mix of enterprises and activities (economic and other) that constitute the nation's economy of social structure. It also supports interactions among smaller groups of uses (e.g., in chat rooms and discussion groups). Information technology stands poised to dramatically transform the way people live, work, and play and the way organizations large and small conduct business. With continued research, development, and deployment, IT systems could enable users to routinely access information of many types (text, images, video, etc.) from any location, participate in continuing education programs from the home or office, shop at their convenience, work from home rather than commute to a central office, consult with medical practitioners remotely, or access government services and receive government benefits electronically.1
As such, IT joins mass transportation and more traditional telecommunications (i.e., telephony, broadcast media) as a foundation for the social interactions that form one basis of society, industry, and commerce (Mitchell, 1996). Those long-standing societal infrastructures—transportation and telecommunications—profoundly affected aspects of society, contributing to the rise of suburbs, the globalization of industry, and the decreasing isolation of political economies. Similarly, networked IT is
increasingly affecting society, as today's debates about topics such as technological literacy and the digital divide attest (Cairncross, 1997). These issues are not necessarily unanticipated—one of the pioneers of the Internet, Leonard Kleinrock, recognized and wrote about some of these challenges in 1974 (Kleinrock, 1974), but they have not been adequately addressed by researchers in technology areas or other fields, and they are growing in importance as IT becomes more pervasive in society.
Today's IT systems put a premium on the explicit consideration of the context (e.g., organizational, societal, or business) in which IT systems are deployed and the organizational structures, human factors, and other types of technology (e.g., transportation and materials transformation) that are involved in completing a certain task. In these social applications of IT, computing, storage, and communications technologies are profoundly influenced by the people involved, the choices they make, and various aspects of human behavior in the design and implementation of the system. Such systems often display a host of other features that make them especially challenging topics for research:
They are often large in scale and high in complexity (see Chapter 3).
They can be geographically distributed and vulnerable to malicious attacks or unintentional errors.
They are often deployed and operated in an environment that is largely uncoordinated.
They have critical requirements for availability and security, with the potential for significant losses (financial, human, or otherwise) if they fail.
They include people and organizations, along with technology, as essential elements.
They are deeply affected by social, economic, and political considerations, such as privacy, productivity, strategic business advantage, national security, poverty, equitable access, and so on.
Their design must take into consideration the human and organizational context in which the systems are deployed and the interactions among people, organizations, and technology.
Although the first four of these characteristics are common to many large-scale IT systems discussed in Chapter 3, the last three characteristics are especially true of social applications and demand special attention. Consider, for example, the systems used in e-commerce or air traffic control. In both of these cases, the interactions among IT, people, and organizational structures are fundamental to system performance. The IT is placed in an existing social and organizational environment in an attempt to improve quality, productivity, speed, and other performance attributes.2
Overwhelmingly, the most important opportunities lie not in simply automating existing applications, but rather in rethinking and remolding the structure and organization of the business process to reflect the best uses of IT and in redesigning and remolding the technology to make it most valuable in its (rethought) application context. The challenge is to reinvent both the application and the supporting technology to make the combination of technology and applications effective. In business, this is often called reengineering or transforming a business process, but the concept applies to the full range of social applications of IT.
Transformation requires a rethinking of the entire sociotechnical system, not just the IT portion. Because most social applications involve individuals, brought together in organizations, and technologies that relate to the movement or alteration of materials or other physical items, a critical part of such a transformation is the identification of the capabilities that each element—people, IT, and other technologies —most beneficially contribute to the process, and the determination of how these elements can best work together. These processes are often constrained by complex social and regulatory issues and must take into account a number of nontechnical factors, such as capital budgets, work rules, skill sets, and administrative organizations. Nevertheless, the creative use of new IT capabilities can result in new, transforming applications. E-commerce, for example, has given rise to electronic auctions. The auction concept is not new, of course; what is new is its application to the selling of common goods and services, both new and used, and the participation of large numbers of buyers and sellers in an electronic marketplace, with new ways for individuals to research price and value and new ways to negotiate. Another example is air traffic control. Advances in technology have produced a fundamental change in the way these systems are conceptualized and designed, moving away from a centralized command-and-control model that controls all aspects of an aircraft's flight plan to a system known as free flight, which will give individual pilots greater autonomy—and more information on which to base judgments (Wald, 2000).
The changing nature of work is broadening the contexts in which IT must operate. Employees are no longer expected to sit at a single computer in an assigned office to complete their work. Even a simple application to enter employee expense reports must be accessible via many different devices in different locations: a desktop computer at work, a desktop computer at home, a laptop computer or handheld devices when traveling, and perhaps even a wireless phone. If an organization cannot offer remote access to IT services, it may limit the effectiveness of its staff. As IT systems are used to manage more aspects of a business, the properties of the IT system and the behavior of employees become more tightly interlinked.
IT-inspired transformations occur at all levels within organizations, as is evidenced in, for example, the flattening of traditional hierarchies, globalization of many business organizations and activities, and emergence of new classes of nomadic workers who are not even assigned permanent offices. As these trends demonstrate, the design of organizations is a matter of both people and IT, leveraging the strongest capabilities of each (Walton, 1989). New methodologies are needed for the design of enterprise applications that more deeply integrate the organizational design and the IT system design. Such methodologies can be developed only through collaborative research in the disciplines concerned, including technology, industrial engineering, business, psychology, and others. The challenges are compounded by the reality that all the information systems that must work together in support of an organization or society cannot be developed at once; rather, new elements are continually added to an existing mix of legacy technologies and applications. Capricious organizational requirements, particularly in a competitive business context, add another dimension to the challenge that has proven difficult to overcome.
Of course, IT has always been designed and used in one context or another. Even traditional applications (e.g., word processing) that could be used on individual computers without networking or pursued independently of other applications have had important interactions with job definitions and human relationships within organizations. 3 Social scientists studying such applications have reported on changes in status, hierarchy, work flow, job design, job satisfaction, productivity, and so on, all of which have contributed to ideas for enhancing early applications and evolving new ones.4 Some of these ideas have contributed to computer science in arenas such as human-computer interaction (HCI, which includes the design of interfaces between people and systems and the design of systems for computer-supported cooperative work), but compared to the opportunities emerging now, those instances of interdisciplinary research are too few and too isolated.5 The evolution of HCI is a promising indicator that progress is possible if social applications are addressed through IT research that draws on multiple disciplines: the subdisciplines “form intertwined roots in computer graphics, operating systems, human factors, ergonomics, industrial engineering, cognitive psychology, and the systems part of computer science” and draw from “supporting knowledge on both the machine and the human side” (ACM, 1992).
A much greater degree of interaction between IT applications and context is now possible.6 Interest in optimizing that interaction—while addressing issues of complexity and scale—creates an imperative for explicit and substantial attention to context in application design and
implementation, which, in turn, implies explicit and substantial attention to the behavior of people as individuals and as members of groups.
Great opportunities await progress in the sociotechnical systems that underlie IT, because such progress would enable far greater capabilities than have yet been implemented in all existing feature sets and their combinations. The IT itself can be used more effectively —and it can be combined better with people and their activities as they work, live, and play. The challenge is twofold: (1) to reinvent social applications to improve the combinations of behavior and IT with the aim of producing better economic and social outcomes and (2) to invent new social applications that enhance the economy, culture, or quality of life.7 Both processes build on past experience with IT, which has demonstrated that tasks are typically automated directly at first and subsequently reconceptualized or reinvented to take better advantage of new technology. Both processes focus on the role of individual users and organizations as major players in social applications.
RESEARCH CHALLENGES IN SOCIAL APPLICATIONS OF INFORMATION TECHNOLOGY
Social applications of IT can motivate research on a range of questions as broad as the applications themselves, with the questions reflecting the particular circumstances in which specific applications are deployed. Some social applications are associated with a given industry or industry sector (e.g., online stock trading systems or flight reservations systems), some cut across the economy (e.g., business-to-business e-commerce systems), and others are specific to a particular organization or function. Despite the fact that it is motivated by applications, the research that results from an examination of social applications can be highly fundamental, in that it requires investigations into, and the development of, basic IT capabilities that are widely applicable to a wide range of systems. The research also tends to be highly interdisciplinary, drawing on the expertise of people in the IT and social science communities, as well as end users who understand the way systems are used in different industries and functions.
Despite their diversity, the social applications of IT tend to have in common a set of elements that support (1) group interaction, (2) knowledge management, (3) commerce, and (4) control and coordination. Although some social applications of IT emphasize one or another of these elements, they usually emphasize at least two (Box 4.1). Each of these types of functionality presents a set of interesting research challenges. An examination of each one will provide a sampling of the types of problems
Social Applications of Information Technology: Examples and Features
that could be addressed if social applications were to play a more significant role in motivating IT research.
An important feature of social applications and their context is that they involve people as members of groups. With networking, computing moves from enhancing the productivity of individuals in tasks they perform alone to supporting the needs and enhancing the productivity or social interactions of groups of people, or helping people find other, like-minded people—a class of applications called group applications in this
report. Group-based elements support activities, including interaction or collaboration, among groups of individual users. Individuals and groups can now publish and manage information, literally on a global scale, as illustrated by many of the well-known Internet applications, such as electronic mail (e-mail), discussion forums, and the World Wide Web. But the Web affords comparatively static, passive information sharing with relatively little accommodation for variation in individual capabilities or preferences. To improve on this technology—and support a wider range of interactions—researchers must delve into group dynamics and interaction, human learning and cognition, human impairments, and other variations on these themes and how they might be supported by IT. Such research depends on insight from psychology and sociology as well as computer science and electrical engineering. It also depends on insight from specific domains that may shape real-world contexts: for example, different requirements will be associated with groups engaged in routine teaching and learning as opposed to groups collaborating on responses to natural disasters, when one can expect extreme variations in available technology, skill sets, responsibilities, and work environments and where crisis conditions affect needs for the type, delivery rate, and comprehensibility of information gathered, analyzed, and shared (CSTB, 1996, 1999).
Knowledge-based elements support the capture, retrieval, and manipulation of knowledge, typically drawing on massive collections of information. Although more and more data are being generated or recorded in networked computers, finding essential information is increasingly difficult. As the nation shifts from an industrial to an information economy (Shapiro and Varian, 1998), the role of physical assets as a source of competitive advantage is diminishing. To see the role that knowledge is playing in the economy, one need look no further than stock market valuations of Internet start-ups that have negligible physical assets but considerable intellectual property. The acquisition or discovery of knowledge (which is derived from information), plus the strategic management and exploitation of that knowledge—a process called knowledge management—are therefore an increasing focus of many companies, both new and established (O'Leary, 1998).8 The goal is to be able to find, understand, and use the massive amounts of information and knowledge that reside within an enterprise.
Tools for searching for information remain frustratingly poor. While it may seem easy enough for people to express their information needs to one another, computer retrieval techniques are unable to filter out a large number of useless search results. To counter the technology shortcomings,
organizations today manage knowledge using a combination of people and computers. Much as a library contains books that record information, as well as librarians to organize and index it, computers are used to store and transmit information but are able to organize and index only data that have a precise, logical structure. Managing knowledge that is encoded as expository text is largely beyond today's IT capabilities.
Historically, IT contributed to information management through systems for collecting and storing information (i.e., databases), finding and retrieving it (i.e., information retrieval), and processing online transactions. These systems provided experience with the types of functions that people now want to extend, combine, and enhance in new and more powerful ways. Data warehousing, which captures a historical record of an entire enterprise's transactions, and data mining, which attempts to analyze data to identify hidden trends and correlations, represent the current state of the art in knowledge discovery and management. For the most part, knowledge discovery and management strain current IT.
Distributed transaction databases with properties such as automatic load balancing and historical archiving have been suggested, but they are beyond the current state of the art.9 In addition, end users are beginning to recognize that knowledge management projects require as much social science as computer science because the systems must serve the evolving mission needs of their users. According to some industry analysts, data warehousing projects fail more often for organizational reasons than for technical ones (Deck, 1999).
Commerce-based elements support the interaction of organizations (including businesses, government entities, and universities) with other organizations and with individuals, whether consumers, citizens and tax-payers, or students. Today's most obvious example is e-commerce, the buying and selling of goods and services, especially among organizations. Business-to-business applications of IT have expanded dramatically from straightforward replacement of paper documentation (such as purchase orders and invoices) in electronic data interchange and electronic bank-to-bank wiring of funds, both of which have been in use for some time. In the emerging model, IT is integrated into all business-to-business operational activities except for the flow of material goods. This new e-commerce (Keen and Balance, 1997) includes activities such as electronic money management (direct transfers of money in electronic forms between businesses), electronic business logistics (coordination of suppliers and customers) and supply chain management (integration of business processes across businesses with supplier/customer relationships).
Most of these activities also apply to nonbusiness organizations, such as government entities involved in procurement, tax collection, licensing fees, and so on.
The development of Internet-based e-commerce has important implications for IT and related research. First, it has amplified the role of information about individuals as an element of business strategy and therefore of knowledge management. Now IT is being used to collect and analyze information about individuals as actual or potential customers—information that can be used for product design, marketing, and customer service. The result is a predictable tension between those interested in the commercial exploitation of information about individuals and those concerned about protecting privacy (Diffie and Landau, 1997). New questions are being asked: Who knows what about whom? Who owns information about individuals? Who is authorized to use that information and for what purposes? These concerns are related to those surrounding the protection of intellectual property in a digital environment.10 Second, as illustrated by the growing attention to privacy as well as the concerns about reliability, dependability, and trustworthiness discussed in Chapter 3, government entities are scrutinizing the nature and use of IT in e-commerce. The result may well involve a combination of voluntary industry actions and government-mandated actions to promote or avoid certain uses of IT, which would have implications for system design and implementation constraints. One certain result would be a further change in the nature and impact of government, which—like businesses and other organizations—is itself affected by the use of IT.11
Finally, e-commerce is redefining the business processes that span traditional administrative and organizational boundaries (Davenport, 1993) and altering the relationships among organizations.12 Even before the commercialization of the Internet, industry witnessed a strong trend away from vertical integration and toward more specialized or horizontally diversified firms, driven in part by lower coordination and transaction costs enabled by IT. There is speculation about the prospects for consolidation and concentration. The equilibrium boundary of a firm is largely set by the relationship between internal and external transaction and coordination costs, and those costs are being profoundly influenced by IT (Grenier and Metes, 1996). Furthermore, the judicious application of new IT can greatly influence these boundaries and the efficiency of the economy as a whole—another area deserving of research.
Coordination and Control
Control- or coordination-based features support the coordination of large numbers of distributed elements—often a combination of techno-
logical, informational, and human—in achieving the reliable operation of a large-scale system. They are especially prevalent in infrastructure systems that perform critical support functions for society as a whole, such as the systems used in utilities (electric, gas, telephone, and networking), finance (electronic funds transfer and auction markets), and transportation (air and rail networks). These systems place extreme requirements on features such as reliability and trustworthiness (see Chapter 3). Many require real-time operation and decision making, some of which will be done in consultation with users and some of which must occur automatically. The specific needs, and the ways in which these needs are met, are closely tied to the nature of the application. For example, security practices limiting access to a system may need to incorporate different override features depending, for example, on whether the system is being used to transfer funds between banks or to allow a doctor emergency access to a patient's medical record.
Despite the diversity in social applications and the IT systems that support them, there are common elements, and they can motivate research that could inform the development of all such systems. Obviously, the broader the range of applications that benefit from particular technological and methodological innovations, the greater the overall value of the research. This is an argument for focusing a large portion of IT research on highly generic problems, such as the following:
Social trust—How can people communicating with strangers in electronic communities know who is trustworthy and who is not?
Coordinating expertise—How can systems and procedures make it easier to elicit and coordinate the expertise of individuals, whether they are doing system development or participating in discussion groups?
Personal privacy and identity—How can systems and procedures make it easier for people to disclose as much as, but no more than, they wish about their identity and personal information in online contexts?
Agency—How can people better instruct IT systems to make decisions on their behalf, e.g., to coordinate calendars and to process and route messages (such as e-mail, faxes, and voice messages)?
Community and collaboration—How can IT better support the activities of groups of individuals seeking to communicate and collaborate more effectively?
Knowledge content—How can the collection of, access to, and extraction of value from vast amounts of information be made more effective?
How can these tasks be better distributed among people and systems of different types?
Business processes—How are repetitive, ongoing activities that meet a business goal best organized around technology, and how is technology best organized around such processes? Can common elements or strategies be used?
Security—How can researchers rethink IT to reduce the vulnerability of the infrastructure, applications, and people to various risks and system failures?
Operations and administration—How can IT be designed for enhanced operation and administration, particularly in a world in which many applications need to interoperate with other applications and across organizational and political boundaries?
Development methodologies—How can software applications be developed more effectively, with the goal of improving outcomes in terms of meeting application needs for functionality, interoperability with other applications, and the flexibility to change in the future?
All of these issues are common to a number of different applications, and all are receiving attention today by researchers in disciplines such as computer science and engineering, information management and science, the social sciences (particularly economics, psychology, and sociology), and business. However, this attention often has too narrow a disciplinary focus. Research in these areas would benefit from a multidisciplinary approach because it pays more attention to application contexts and could define new issues and requirements to inform the research.
In addition to research in generic areas of concern such as those listed above, IT research can address more specific application contexts as well. The purpose, again, is to inform the research on IT by considering realistic needs and contexts, with the goal of molding the technology to be more powerful, effective, and secure.13 There are important reasons for doing this. First, any particular area of technological research is more critical to some applications than to others. Some applications stress a particular facet of technology. By identifying and studying the applications that are most challenging, the IT research outcomes can have a broader and greater impact. Second, researchers pursuing generic technological concerns tend to ignore interactions with other concerns or miss opportunities to address different areas of concern with more holistic solutions. The only practical way to appreciate a full range of issues and how they interact with and buttress one another is to look at the whole problem in specific applica-
tion contexts. Finally, as the application domain of research narrows, the impact of technology improvements to that application domain are likely to be greater.
As an example, consider the IT research objective “greater flexibility to meet changing needs.” Can flexibility be addressed generically? Certainly, as outlined in Chapter 3, and indeed it is worthwhile to search for general approaches and principles. But it is also worthwhile to first identify more specifically what sort of flexibility is needed, as informed by particular applications domains. Does flexibility imply the merging of information systems when firms merge? Does it mean adding new applications to a mix and ensuring that those new applications are interoperable with the old? Does it mean adding features or modifying the features of existing applications? Does it mean layering new applications on top of old ones? Are there other options to explore? To what extent is the definition of flexibility itself flexible? By examining particular application domains, researchers can answer such questions with greater specificity, and the results may be more immediately useful. At the same time, of course, the results become less generally applicable; yet it is often possible to generalize from the specific case studied to a broader range of similar cases.
An essential part of research on the social applications of IT is building and deploying experimental systems. Often the utility of an application cannot be predicted by its designers or by past experience. For many IT applications, the task is more like one of industrial product design than business process reengineering. Standards for IT research need to acknowledge emerging social effects, not focus exclusively on the technology. The World Wide Web, for example, is based on extant technologies for document management and hypertext, but the unique combination of these technologies addressed a social need (to publish and communicate) with a power that could not have been predicted from previous experience with the technologies.
CONDUCTING RESEARCH ON SOCIAL APPLICATIONS
A Plausible Approach
Research that addresses the social applications of IT would need to target applications that serve groups of users, organizations (including government, businesses, and others), the citizenry at large, and critical infrastructure systems. Examples include the following:
Groups—Collaboration, political activism, battlefield management, community development;
Organizations—Tax collection, census collection, military logistics, e-commerce;
Citizens—The political process (polling and voting, citizen involvement in decision making), continuing education, entertainment, cultural opportunities; and
Critical infrastructure—Transportation and communication networks, intelligence gathering, nuclear power, and financial markets.
The research would not only address today's nascent social applications (such as e-commerce) but also explore visionary new possibilities that are unimagined today or even unimaginable owing to their scale and complexity. Much of this research would aim for visionary and revolutionary advances through the creative application of technology. Some research would work from the bottom up, envisioning the potential capabilities of technologies and identifying new opportunities that exploit those capabilities. Other research would work from the top down, categorizing the major challenges experienced in the world today or tomorrow that might be overcome by the creative application of existing and new technology. The research would aim to enhance the positive effects of IT while identifying and minimizing its negative effects.14
Similar to systems research, which needs to pursue both case research and methodology research (as discussed in Chapter 3), research on social applications would need to pursue both case research and generic research. Case research would identify target application areas such as “digital democracy” (i.e., public participation in collective decision making) and address the role IT could play in them. Such research would also seek new mechanisms for the political process and supporting technologies. Generic research addresses issues such as group dynamics, decision making, collaboration, enjoyment and satisfaction, and competition and looks at how they apply to and can be mediated in social applications. Generic research could also address how technology can be molded to help in balancing economic benefits to organizations against the rights or quality of life of individuals. An important example of a generic research topic is privacy; research would attempt to define what privacy entails in different contexts and identify technological means to control the degree of privacy as a function of circumstance and objectives.
Outcomes of the research into the social applications of IT could be evaluated in terms of criteria such as (1) the extent to which society and organizations benefit more from technological progress than they would otherwise, with respect to whatever dimensions make sense in context (e.g., productivity, quality, or effectiveness), (2) the extent to which the negative consequences of technological progress are avoided or mitigated, and (3) the extent to which technology is influenced by beneficial long-
term, visionary ideas rather than being simply the result of incremental improvements and near-term commercial exploitation.
Research of the sort suggested in this chapter would help society reap the rewards of “learning by doing” and would accelerate that process. It would help researchers recognize patterns in their results and develop new theories and would enable them to select better research targets and exploit their work more effectively. The new approach to research would take nothing away from the tried-and-true work that continues to offer many benefits; it would be an additional, complementary asset in the overall production system.
New Research Teams
The tight linkages between the technical and nontechnical components of sociotechnical systems mean that effective research will have to draw on a broad range of constituencies, much as work in civil engineering is influenced by architects, urban planners, and transportation engineers (Box 4.2). Technology researchers, social scientists, and domain experts will have to work together to identify and solve problems related to the social applications of IT.
To date, research on the social applications of IT has tended to fall into one of four categories:
Research on IT systems addressing generic issues of technology, informed (it is hoped) by a knowledge of application needs and addressing a broad swath of application contexts rather than focusing on a single context (examples include many of the issues mentioned in Chapter 3, such as security, administration, and flexibility);
Applications research that addresses a specific context, trying to see how existing technologies can provide value to that application and how that application context (such as an organization) can be reorganized around IT technologies;
Software engineering research that explores generic techniques for developing and deploying new IT-based applications; examples include the processes involved in understanding the application context, translating this understanding into requirements and specifications, implementing the applications, measuring and refining their behavior, deploying the application, training users, and supporting the deployed application and its users; and
Social science research that explores how people undertake different activities and exercise their different physical and mental abilities to engage IT and other resources and research that explores how people adapt the structure of organizations and activities in response to the intro-
duction of IT. This category includes the classical social science disciplines themselves—e.g., psychology, social psychology, sociology, and anthropology.
Social Applications of Technology in the Construction Industry
The interplay among civil engineering, architecture, transportation engineering, and urban planning demonstrates the multiple influences that impinge on social applications of technology, in a setting that many people find somewhat more familiar than the information technology (IT) industry. As such, the construction industry provides a useful illustration of the direction in which IT is heading.
Civil engineers are adept at achieving cost-effective structures based on well-specified plans provided by an architect. The architect is adept at generating those plans by understanding the local conditions and environment and interviewing future occupants of the building about their needs and wishes. But neither of these professionals can work in isolation. Their work relates to both transportation engineering and to land-use planning, and it is not difficult to see that isolated decisions by architects and civil engineers can have unintended consequences in one of those domains, perhaps interrupting traffic patterns or altering the character of a neighborhood. Local interests, which may be expressed in plans for a given structure, can be varied, but they must be aggregated for purposes of transportation and land-use planning decisions.
The construction process is clearly complicated by the involvement of three or four professions with distinct cultures and perspectives. However, this is one arena in which there is a history of interaction and a comparatively mature legal and political environment that impinges on the engineering activity, which—until the recent advent of intelligent transportation systems and the prospect of smart homes, thanks to progress in IT—has itself been comparatively mature and stable for a long time.
Clearly, a great deal of valuable research can be performed in these four categories. Much of the research agenda discussed in Chapter 3 can be cast into the first category, as can the research into component technologies that can and should be pursued (discussed in Chapter 1). Much of the research carried out in social-science-based disciplines, such as information and library sciences and business, can be cast into the second and fourth categories. A modest amount of research (in comparison to the size of the challenges) has been conducted in the third category. It has been pursued most notably by organizations such as the Software Engi-
neering Institute at Carnegie Mellon University and some academic computer science departments.
Each of the four kinds of research has limitations if it becomes the sole lens through which to examine the social applications of IT. Because social applications of IT link technology so tightly to people and organizations, a more integrative approach is needed. In the technology arena, IT is subject to few physical limits, so it can be molded in many different ways. To consider the technology in isolation is to miss opportunities to mold it in directions that make it more useful. Similarly, because a great deal of research in application areas takes the current observed state of technology as a given, it misses opportunities to think about ways in which the technology could be different that would make it much more useful. One of the great unsolved problems is how, as a part of software application conceptualization and development, to take into account the rich interaction between the application context and the architectural assumptions in the software application, as well as the issues of legacy technology and applications and changing requirements. This last challenge clearly embodies the consideration of technology (as it is and as it could be) in conjunction with a deep understanding of the context.
There is tremendous potential for expanded research agendas that fall between these categories and combine multiple perspectives into a more cohesive, holistic view. Research that examines the impact of IT on society and the economy is certainly needed, but equally important is research on the impact that application contexts have on IT development and software development methodologies as a part of the larger process of application conceptualization and refinement. New science and engineering may arise from such research. The discussion in this report emphasizes research in which the goal is to refine the IT itself, often in conjunction with meeting application needs more effectively. Impact remains an important issue, but there is an explicit goal of modifying the technology to maximize the beneficial impacts and mitigate the adverse ones.
Who is in the best position to make contributions to IT as informed by applications—especially the social applications of IT—and how can research on this topic best be organized? The most effective researchers will have an understanding of the potential of IT, both what might be possible technically as well as what has already been achieved, as well as an understanding of the application context. In general, opportunities for advancing application areas using existing technologies are best addressed by application domain experts who have acquired a fairly substantial understanding of the technology. (There are certainly opportunities for technology experts to participate in this research in consulting roles, but such participation is generally not greeted with enthusiasm.) The use of application
domains to inspire technology is a role predominantly for technology researchers who inform themselves about the characteristics and challenges inherent in particular problem domains. Such knowledge can be gained in at least two ways: (1) by working directly with the end users (individuals and organizations) in that application domain or (2) by collaborating with researchers who have the appropriate domain expertise or expertise in disciplines directly relevant to the domain.
Where the goal is to advance both the technologies and the ways of addressing application domain challenges, interdisciplinary research collaboration is generally needed. Although some degree of collaboration has always been necessary, the need for it is growing quickly, driven by the availability of the Internet, the proliferation of IT applications in consumer and other layperson contexts, and the rapid expansion of social applications of IT generally. These trends also argue for substantially expanding the involvement of end users in setting the research agendas and articulating the goals and constraints of the research, for it is the end users who experience the opportunities and challenges in an immediate and detailed fashion. Although several programs have been established to bring together the diverse set of stakeholders needed to make progress on social applications of IT, it is increasingly apparent that a broader and more substantial effort is needed that uses multiple mechanisms to stimulate the type of research that will make significant progress. There are many practical barriers to the desired intellectual cross-fertilization, but mechanisms exist for overcoming them, as outlined in the next section.
MECHANISMS FOR SOCIAL APPLICATIONS RESEARCH
Participation of End-User Organizations
To date, end-user organizations have relied largely on their vendors to perform long-term research in IT. Vendor research, which emphasizes components, needs to continue unabated if rapid advances in the speed and capacity of IT are to be maintained. However, when it comes to research that affects the uses of IT, the constituencies who understand and benefit from those uses—that is, end-user organizations—must participate. These organizations include (1) companies that are strongly dependent on IT, including firms in almost all industries but especially those in business and financial services and online commerce; (2) suppliers of enterprise software used to automate major processes for corporations, universities, and government agencies; (3) application developers that develop customized software for large companies and government agencies; and (4) systems integrators that integrate major infrastructure
components (hardware and software) to meet the needs of large-scale applications.
End users also include government agencies that depend heavily on IT systems—especially those that have considerable difficulty deploying such systems effectively. This group includes both traditional federal supporters of IT research—the Department of Defense, the National Science Foundation, the Department of Energy, and the National Aeronautics and Space Administration—as well as agencies that have a more limited history of supporting research on IT, such as the Internal Revenue Service, the Social Security Administration, and the Federal Aviation Administration. All parts of the government are increasingly dependent on IT. In FY00, federal agencies were budgeted to spend more than $34 billion on IT systems, up more than $1 billion over FY99.15 Information technology is used both to run the back offices that store data and record transactions and to interact with citizens doing business with the government. Arguably, the funding of basic IT research directly supports the missions of these agencies (Wulf, 1999). If new IT developments can increase the effectiveness of these functions or reduce their cost, then the nation as a whole will benefit.
The greater participation of end-user organizations in IT research could have numerous benefits. First, these organizations can bring substantive intellectual insight to the research process. Research on social applications of IT demands an understanding of the ways in which IT is deployed in different organizations, the missions it supports, the requirements it must meet, and the problems that are experienced. Such insight can come only from end-user organizations that work with such systems routinely and can foresee the ways in which such systems might be used in the future. Recognition of the benefits of researcher access to real systems in situ is not new. The CSTB, for example, first recommended such access in 1989 in a report that provided recommendations for research on complex software systems. The participants at the time noted that the problems identified dated back to the 1960s, and a reading today suggests that many of the same problems persist now in software engineering and related research (see Box 4.3).
Second, end-user organizations could contribute funding, thereby expanding the financial base for research related to IT. Such an expansion is necessary if research motivated by social applications is to grow without crowding out existing research on components. The well-publicized and expensive failures of IT in end-user organizations also make a compelling case for the expansion of research funding from all sources, including the government. Finally, end-user organizations will insist (appropriately) that the research have an identifiable impact on the problems they experience. Thus, their involvement will focus the research community's
Excerpts from Scaling Up: A Research Agenda for Software Engineering
Short-Term Actions: Foster Practitioner and Researcher Interactions
Long-Term Actions: Legitimize Academic Exploration of Large Software Systems
Glean Insights from Behavioral and Managerial Sciences
SOURCE: Computer Science and Technology Board (1989), pp. 19-21.
attention on social applications of IT to a much greater extent than would otherwise be expected.
Engaging end users more directly in IT research will not be easy. Many cultural differences must be bridged, and practical issues in the management of research must be resolved. A sociotechnical systems perspective on any large-scale IT in context requires the identification and
analysis of human and organizational shortcomings as well as technical ones. It is difficult to gain access to organizations, let alone financial support, for projects that are likely to identify human shortcomings (as well as technical ones) and publish them in the open literature. Organizations may pay consultants for such work already, but contractual terms of nondisclosure guarantee that the findings will not be subject to the scrutiny of peer review and will not become general knowledge. Furthermore, because end users and systems integrators have not been actively engaged in IT research, few have any internal research capacity or ability to manage outside research and ensure its quality and relevance. Consequently, they may be unable (at least at first) to participate effectively in the conduct of research. If research is conducted by parties outside the organizations, mechanisms will be needed to ensure that the research will benefit the end-user community and influence actual practices. Researchers themselves will need active and ongoing contacts with end-user organizations to gain hands-on experience with real issues and problems and the ability to validate their ideas empirically in realistic environments.
One approach would be for end-user organizations to pool their resources and fund external research related to their interests, whether at universities or industrial research labs, or in consortia established expressly for this purpose. This approach has been used by IT firms to support research of mutual interest through organizations such as SEMATECH, the Microelectronics and Computer Technology Corporation, the Semiconductor Research Corporation, and others (see Box 4.4). Such arrangements have produced mixed results to date; their success depends on many factors, including the ability of the member organizations to agree on a set of issues to address, the leadership of the consortia, and the quality of the researchers (many are drawn from member companies). When this model is used for research on the social applications of IT, the mechanisms for transferring technology back to the end users could be a problem. There is ample evidence that before they can benefit from research consortia, firms must have their own research and development (R&D) operations in place. Professor Wes Cohen of Carnegie Mellon University has referred to this characteristic as “absorptive capacity.” Consortia in e-commerce or medical informatics, for example, would not succeed if banks or health maintenance organizations did not have their own internal research operations under way.
Accordingly, one element of a larger strategy would be to try to increase the amount of direct R&D conducted by firms that use IT systems. Many an applied research organization leverages external research outcomes—whether funded by that organization or not—for internal benefit. These organizations could be involved in choosing the recipients of research funding. Importantly, they have experience in identifying inter-
Models for Pooling Information Technology Research
Companies in the information technology (IT) industry have at times joined forces to fund research that will aid the entire industry. The examples below could serve as models for end-user organizations to pursue as they increase their participation in IT-related research.
nal problems and challenges worthy of research, and also in identifying individuals who can convey those challenges to external researchers. In other words, they are matchmakers, putting together internal and external resources and facilitating access to internal facilities for empirical research activities. Such organizations can also be vehicles for ensuring that the research project is chosen in consultation with organizations well versed in this process. Government agencies, for example, could enlist the aid of federal agencies that are experienced in funding IT research, such as the NSF, DOE, NASA, and the Defense Advanced Research Projects Agency (DARPA).
The federal government is attempting to forge stronger links between the IT research community and end users in federal agencies. The Digital Government program, for example, uses the NSF as an intermediary to fund IT research addressing the medium- to long-term R&D and experimental deployment needs of federal agencies.16 The program requires at least one government agency to be “significantly involved in defining and executing the research” and requires recipients of grants to integrate into their projects experts in domains that are primarily or exclusively associated with government. Agencies are encouraged to share facilities, data, or personnel with the researchers and/or to provide funding either directly or jointly with the NSF.17 Although the program is new and small (approximately $3 million was allocated to it in FY00), it has succeeded in attracting a number of proposals that involve a range of federal agencies with little or no experience in funding IT research.18 However, many projects reflect the agencies' interests in developing software and systems to meet particular near-term needs and their desire to examine different technological solutions to determine which offers the best combination of features. NSF program managers hope to create a research culture in federal agencies that will promote longer-term research and facilitate the agencies' adoption of new technologies once the research is completed, but this process will take time. Mechanisms may be needed to demonstrate the feasibility and applicability of new technologies to help agencies better evaluate their risks and benefits (Larry Brandt, NSF, “The NSF Digital Government Research Program,” presentation to the committee, June 14, 1999).
Another attempt to strengthen linkages between researchers and end-user organizations is IBM Corporation's First-of-a-Kind (FOAK) program. 19 The FOAK program, initiated in 1995, links IBM's Research Division with its industry solution units (ISUs) to apply innovative technologies to real customer problems. The idea is to do research in the marketplace as well as in IBM's laboratories so that customers can gain access to innovative technologies, researchers can get early marketplace feedback and insight into the applications of new technologies, and the ISUs get solution assets that could be reused to solve other, similar customer problems. By choosing the appropriate customer problem to work on, researchers hope they can generate solutions that are general enough to solve other customer problems not only in the same industry segment but also in other segments.
The FOAK projects are jointly funded, implemented, and adminis-
tered by IBM Research and the ISUs. Starting in 2000, IBM created a new review board for the program that draws members from the Research Division and IBM's Global Industries Division. Managers have also emphasized the need for a lead person who is responsible for driving the resulting innovations into the marketplace. Leaders in several divisions (e.g., Global Services, Lotus, and Sofware) have championed proposals and sponsored project teams in two of the divisions (Research and Global Industries). These teams work with ISUs and business partners, who bring domain expertise and marketing experience to the effort.
Roughly 25 percent of IBM's research budget is dedicated to the FOAK program. Since the program started in late 1995, more than 80 projects have been funded. Of these, more than half have been successfully completed and their outcomes deployed in the marketplace (either as a product or as part of a solution). About 15 percent of the projects were terminated, either because they could not get a customer commitment or failed to achieve key project milestones. As of April 2000, the program had about nine active projects and was driving considerable patent activity. Many researchers were skeptical in the beginning (“What do you mean, work on real customer problems?”), but most have come to view the program as valuable to their long-term research. By engaging early on with real customers in the marketplace, researchers have been able to gain early insight into the value and application of their technology and into future directions for their research. Successfully deployed systems include an Internet-based electronic ticketing system for Swiss Railway, a set of sophisticated pattern-matching algorithms used to discover molecular relationships in the process of new drug development, a continuous speech recognition system for radiological transcriptions at Massachusetts General Hospital and the Memorial Sloan-Kettering Cancer Institute, and a mail analyzer that allows customers to automatically queue problem reports.
Industry Internships and Sabbaticals
Another means of increasing the linkages between industry end users and academic researchers is industry internships and sabbaticals. Such programs can help ensure that academic researchers develop an understanding of the challenges faced by end-user organizations (and vendors) in designing, developing, deploying, and operating IT systems. Of course, many students who graduate with advanced degrees in IT go to work in industry or government IT settings; in that sense, the university perspective is introduced into industry every time a student is hired. Internships in industry before students graduate could add real value, particularly if students select thesis topics that focus on some aspect of a real-world IT
problem or social application. Work in the social applications of IT, in particular, would benefit from greater cross-fertilization between industry and academia, because it could provide academic researchers with access to large-scale computing and communications artifacts (i.e., operating physical systems) and the individuals responsible for them.
Similarly, exchange programs in which faculty members work in industry or government for a year and industry or government IT professionals work at universities for a year would help build a cadre of experts who are comfortable in both worlds. Alumni of such exchange programs could go on to play a valuable role in both research centers and interdisciplinary IT research groups. Equally important, both they and the veterans of student internships would form a core of individuals who could apply new research findings by designing and operating improved large-scale systems and social applications.
A number of IT-related internship and exchange programs exist, but most place students and faculty members in vendor organizations. The expansion of such programs to include end-user organizations might be facilitated by federal incentives. Because few end-user organizations (or systems integrators) have large research operations, they do not have a tradition of supporting interns in IT, nor do students view internships with such organizations as furthering their research careers. Seed funding for such a program through a federal agency such as NSF could encourage such internships and, if the program is successful, could convince industry to contribute funding as well. Universities might need to be persuaded that there is value in such programs, which would take students away from their faculty advisors for some period of time and could be viewed as delaying the completion of their studies. At the same time, however, such internships could give students valuable research experience and help them to select dissertation projects (or thesis topics). If successful, such internships and sabbaticals could also serve to demonstrate the value of IT research to end-user organizations, prompting them to support additional university-based research, benefiting both communities.
Industry experts could also be brought into the university system. Many universities tap into the expertise in local industry by hiring industry employees as adjunct faculty. Although adjunct faculty members do not typically participate in research projects, they could introduce new ideas to students and other faculty members alike. The Massachusetts Institute of Technology (MIT) has begun to experiment with another type of program to bring industry expertise into academia. It has appointed a small number of “professors of the practice,” who teach and conduct research full-or part-time. These individuals are not expected to have publication records as long as those of their academic counterparts but
are instead expected to have a significant record of accomplishments. The program is still new, and MIT has only a small number of such professors, but they add breadth and experience to the school's engineering programs.
Interdisciplinary Research in Academia
If research on the social applications of IT is to advance, then interdisciplinary research will be needed that involves participants and expertise from a wide range of academic disciplines. The work needs to involve not only computer scientists, engineers, and software experts but also business professors and organizational theorists who understand the relationship between IT and the organizational structures within which it is embedded, the human side of complex technical systems, and the market aspects of different social applications. The research teams need to include social scientists who can evaluate the impact of IT on individuals, families, organizations, and society and who understand the human-centric nature of computing. They could include specialists in particular application areas such as health care, manufacturing, finance, and e-commerce.
Researchers in business, economics, and the social sciences have already made numerous advances in understanding IT as it is used in a range of social applications (Box 4.5 and Box 4.6). Several opportunities exist for stepping up the involvement of experts in business, economics, social science, and law in research pertaining to IT. Nontraditional research mechanisms may be needed that will encourage the participation of end-user organizations in research, broaden the outlook of IT researchers, and/or overcome disciplinary boundaries in universities. The management of interdisciplinary research collaborations generates its own set of issues: technologists and social scientists have different vocabularies, methodologies, time perspectives, standards of evidence, and so on. Such differences need to be bridged if collaborations are to be effective.
Interdisciplinary research can be conducted in one of three ways. Individuals can broaden their own expertise: technologists can become increasingly facile with the uses of the technologies and the larger system (including sociotechnical system) contexts within which they are embedded, and experts in sociotechnical system contexts become more facile with technology. Alternatively, experts in the system contexts can collaborate more extensively and effectively with technologists. Or, new professions can arise to mediate between IT and its uses: this occurred in health care when the field of medical informatics emerged to help bridge the gap between IT and medicine (Box 4.7) and, even earlier, in civil engineering when an entire profession (building and landscape architecture) was established to deal with application, societal, and aesthetic
Examples of Business Research Related to Information Technology
SOURCE: Thomas Malone, Sloan School of Management, MIT, in a presentationat the National Academy of Engineering workshop “How Can AcademicResearch Best Contribute to Network Systems and Communications?” Washington, D.C., October 30, 1998.
issues. Each of these approaches has its strengths and weaknesses. The first approach, for example, effectively creates visionaries and leaders in a field but does not enable large-scale collaborations. The second enables a properly balanced team to be assembled but creates problems of coordination and management and makes it difficult to create a shared vision. The third approach can be effective in establishing a long-term capability in an interdisciplinary area, but these areas can become somewhat separated from their parent disciplines. All three of these approaches combined can make a significant difference.
Examples of Economics and Other Social Science Research That Has Contributed to Information Technology Development
SOURCES: Marvin Sirbu, Carnegie Mellon University, presentation atthe National Academy of Engineering workshop “How Can Academic ResearchBest Contribute to Network Systems and Communications,” Washington,D.C., October 30, 1998; Lee Sproull, New York University, personalcommunications dated August 22, 1999, and April 11, 2000.
Lessons from Medical Informatics
Insight into the challenges of interdisciplinary research related to information technology (IT) can be drawn from the field of medical informatics, which deals explicitly with linking the field of medicine and IT. A handful of universities and medical schools have established programs in medical informatics that combine training in both medicine and IT.1 Graduates of these programs can be found in academic medical centers, health-related Internet start-ups, and other organizations. Some work in the IT departments of large health care centers. The number of health care organizations that have achieved a critical mass in this field is small.2 All of these organizations employ physicians funded by IT payroll dollars who work with mainstream IT to improve its application to the core domain of the organization (medicine), with a heavy emphasis on the academic aspects of the intersection between IT and the user domain and an interest in experimentation, prototyping, and evaluation of impact. Results are published in academic journals and forums.
Several factors have contributed to the success of these programs. The first is that academic medical centers have a tradition and culture of research. The National Library of Medicine (NLM) funds training and research, and this funding continues over several years, leading to a measure of program stability. The programs funded by the NLM reside in the same organization as the IT function does (the university medical center), so there is organizational colocation. It is clear to the organizations that many complex initiatives require a deep knowledge of medicine to be solved correctly. The IT function (in particular, the chief information officer) must be comfortable with the academic component and the complex medical domain discussion that medical informatics brings to the table. Organizational mechanisms have been established to assimilate the research functions, such as new or revised pay grades, project approaches to involving researchers on teams, and tactics to develop budgets.
These factors are present in only one or two dozen out of a total of 5,000 health care organizations. At times, the linkage between medical informatics and the IT organization does not work very well. For instance, medical informatics programs do not solve all instances of IT staff failing to work well with end users. Medical informatics staff members are involved in a minority of a health care organization 's projects; however, they are involved in the projects in which the domain (medicine) places the greatest design and operational stresses on implemented systems. One such system is computerized medical records, which are exceptionally complex to design and present numerous work-flow challenges.
1 Examples of these programs can be found at Columbia University, Stanford University, and the University of Utah.
2 In 2000, no more than two dozen health care organizations had significant medical informatics efforts. These included Partners Healthcare System in Boston, Children's Hospital in Boston, Vanderbilt University, Intermountain Healthcare in Salt Lake City, Columbia Presbyterian Medical Center in New York, and the University of Michigan.
SOURCE: John Glaser, chief information officer, Partners HealthcareSystem, Boston, Massachusetts, personal communication, July 13, 1999.
Interdisciplinary research related to IT is ramping up in universities, albeit slowly. An increasing number of active researchers in the information systems field list schools of management and business administration as their primary affiliation. Other important communities of researchers that are often overlooked are those in departments of communication and schools of education. All of these groups include individuals working on both fundamental and targeted research motivated by social applications of IT. In addition, an increasing number of faculty members with computer science degrees are appearing in other departments across campuses, particularly in information sciences and business. They are involving themselves in application areas such as digital libraries and e-commerce, focusing on both the social and technical aspects and collaborating with economists and legal specialists. A movement in human-centered computing, involving collaboration with disciplines such as sociology and psychology, is flowering in academia, albeit at a slower pace than similar efforts in the United Kingdom and Scandinavia.
A major obstacle to any increase in interdisciplinary work is the strong disciplinary orientation of many universities. Academic research is typically reviewed from the perspective of a particular discipline, similar to the way in which grant proposals are peer reviewed. The reviewers tend to be disciplinary experts and rarely reflect the interdisciplinary nature of the research. Efforts to establish interdisciplinary programs within disciplinary boundaries and to evaluate those programs often meet with criticism because they are viewed through the lens of a single discipline.20 As a result, faculty members are often discouraged from pursuing interdisciplinary research. This is especially true early in a professional career, when gaining tenure is a major goal and a faculty member feels particularly vulnerable to peer pressure. Unfortunately, early habits often persist well into a career.
Traditional disciplinary work is reinforced by traditional disciplinary culture. Computer science, like any other field, has its own sets of terms, attitudes, norms, and customs with regard to what constitutes research and how to conduct it. Computer scientists have even found it difficult to collaborate with electrical engineers, although in general they collaborate more easily with researchers from engineering and the physical sciences than with those from the social and life sciences. Reasons for the difficulties in collaboration include factors such as project definition, laboratory orientation, availability of physical infrastructure and professional staff, teaching loads, amount of funding for a project, and so on. More subtle obstacles are attitudes about the relative merits of different fields, with
each field manifesting its own flavor of chauvinism. These difficult-to-quantify realities complicate the establishment of collaborative relationships and projects.21
One solution to the problem has been the establishment of new schools, divisions, or departments (typically drawing on existing resources) within universities to encourage interdisciplinary research and education related to IT and its social, political, and economic contexts. For example, Carnegie Mellon University has long promoted interdisciplinary research by establishing separate departments and programs to pursue them. Its Department of Engineering and Public Policy—and others like it at MIT, Washington University, and elsewhere—have examined issues raised by the intersection of IT and public policy for several decades.22 Many faculty members have joint appointments in these organizations and traditional academic departments. In recent years, a number of universities have transformed schools of library science into broader institutions that address issues linking IT and social applications (Table 4.1). These schools tend to draw faculty from a number of departments, including computer science, economics, social sciences, law, and business, but they can hire and promote faculty on the basis of their interdisciplinary work. In such an environment, a faculty can be built with backgrounds in a diversity of core disciplines, and the evaluation process can relatively easily be appreciative of contributions that cross disciplinary boundaries, in no small part because faculty who join such units have the appropriate vision and orientation. Many of these schools are as yet untested in terms of their abilities to sustain quality research of the type needed to make progress on social applications. They nevertheless exemplify the kinds of efforts that will be needed to make progress in this area. They also present more new opportunities for educating students in interdisciplinary research areas. By creating institutions with permanently assigned faculty, these schools and departments can develop curricula and teach classes more effectively than is generally possible if faculty are scattered throughout multiple academic departments.
Limited funding is another obstacle to greater interdisciplinary work at universities on the social applications of IT. Although there are many notable examples of university researchers who followed a vision and were able to sell their visions to potential funding sources, the much more common model is to shape the research in the directions favored by existing funding sources. Thus, as the funding goes, so (largely) goes the research agenda. The gross disparity between funding available for research in social science and that in computer science today does not bode well for prospects of new social-science-based activity.
Federal support for university research on the social and economic aspects of IT (one component of the larger research agenda motivated by
TABLE 4.1 Recently Established or Expanded University Programs on the Sociotechnical Aspects of Information Technology
School or Department
School of Information Studies
Information behavior, information retrieval, information policy, information and change, information management
University of California at Berkeley
School of Information Management and Systems
Information storage and retrieval, human factors, information policy (e.g., economics and intellectual property), management, networked information systems
University of California at Los Angeles
School of Education and Information Studies
Information technology and institutional change, technology and privacy, genre theory, linguistic aspects of computing, Internet culture, information search and retrieval, social effects of information technology, digital libraries, information policy, user-centered design
University of Illinois at Urbana-Champaign
School of Library and Information Science
Community architectures for networked information systems, information retrieval, computer-supported cooperative work, agents and multiagents, organization theory, information technology (e.g., artificial intelligence, distributed systems, groupware, human-computer interaction), information policy and public policy
University of Michigan
School of Information
Community technology, information economy, electronic work, digital libraries, archives and record management, human-computer interaction, information economics, management, and policy, library and information science
social applications) has come largely from NSF. The Computer and Information Science and Engineering (CISE) directorate's program in Computation and Social Systems (CSS; formerly the Information Technology and Organizations program) supports work in two related areas: (1) the integration, sustainable use, and impacts of IT in groups, organizations, communities, and societies and (2) theories and technologies for reasoning, decision making, interaction, and collaboration in groups, organizations, communities, and societies. For many years, this program has been the NSF's main source of support for truly interdisciplinary sociotechnical IT systems research.23 Its awards have routinely gone to research groups that involve technical specialists from computer science or related science and engineering disciplines, as well as social scientists with deep interests in the technology and its social implications. When groups were made up solely of social scientists, the proposals still had to be reviewed by people with technical expertise.
For the most part, these awards tend to be small, providing support for one or two researchers over the course of 3 years or so (Table 4.2), and the total amount of annual funding has also been small, in the range of a few million dollars. As a result, only a small subset of the issues that need to be addressed is represented in ongoing research projects, and the technological aspects of the program are directed at component technologies (e.g., artificial intelligence and agents) rather than the large-scale social applications. More importantly, the smallness of the grants precludes the ability to bring together many experts from different disciplines (e.g., IT and the social sciences). The CSS program is expected to grow in the near future and fund a variety of work on social applications of IT (Box 4.8), but it is not clear how quickly this will occur or how large the projects will be or whether the program will begin to fund larger teams of researchers.24
TABLE 4.2 Recent Awards by the National Science Foundation's Computation and Social System Program
Estimated Size of Grant ($)
Approximate Duration (years)
Studies of Decision Making in Complex, Dynamic Environments
Algorithmic Issues in Collaborative Filtering
Coordination: Integrating Organizational Style with Environmental Characteristics
Seeing Is Believing: The Value of Video for Remote Interpersonal Connections
The Design of Reputation Systems
Decision-Making in the Context of Commitments to Team Activity
Design of Time Cognizant Electronic Brokerages
SOURCE: National Science Foundation awards database. Available onlineat <www.nsf.gov>.
Representative Topics for Future Support by the National Science Foundation's Computation and Social Systems Program
SOURCE: National Science Foundation awards database. Available onlineat <www.nsf.gov>.
On a larger scale, the NSF has attempted to address the social applications of IT from an institutionwide (cross-directorate) perspective. In 1997, it launched the Knowledge and Distributed Intelligence (KDI) initiative (see Box 4.9), a bold experiment in fostering cross-disciplinary research inspired by social applications of IT that originated within the CISE directorate and the Social, Behavioral, and Economic Sciences (SBE) directorate. Its existence was brief, in part because of programmatic uncertainty both within NSF and externally in the research community. That uncertainty related to the goals and approaches associated with the initiative and the intellectual and practical challenges of pursuing cross-disciplinary research.
Knowledge and Distributed Intelligence
[Knowledge and distributed intelligence (KDI)] activities aim to improve our ability to discover, collect, represent, transmit, and apply information, thereby improving the way we conduct research and education in science and engineering. These efforts promise to change how we learn and create, how we work, and how we live. . . .The objective is to create networked systems that can make all kinds of knowledge available to anyone, located anywhere, anytime.
—National Science Foundation KDI brochure
The evolution of KDI as a broad theme within the National Science Foundation reflects the integration of multiple streams of research and development drawing on many scientific and engineering disciplines. The components are enormously varied, including, for example, research in computational biology, computer networks and communications, high-performance computing, database management and information retrieval, mathematical modeling and simulation, artificial intelligence, human learning and cognition, science, mathematics and engineering education, geospatial information systems, and science and engineering indicators. These components are combined into programs of varying scope and scale.
The KDI activity aims at a new level of intellectual coalescence. It recognizes the progress made to date in developing and deploying information technology across science and engineering research, and it recognizes that challenges for the future include assuring that such technological support can be used and is useful. It attempts to move beyond mere access to information and to develop methods to intelligently absorb, refine, and analyze information to glean useful knowledge. Organizationally, KDI is a cross-directorate, cross-disciplinary effort with three core components:
In addition to the core components, there are currently six KDI-related initiatives in specific technological or content domains that are evolving in cooperation with other federal agencies and private-sector organizations: (1) universal access, (2) digital libraries initiative, (3) Next Generation Internet, (4) integrated spatial information systems, (5) functional genomics, and (6) digital government. These initiatives expressly address issues of how to design and develop technologies that can be used more effectively and by more people. They reflect experiences with early information technology in a variety of contexts as well as ambitious objectives for progress.
SOURCE: CSTB (1998b).
The NSF's support for work related to social applications of IT will probably expand in coming years as IT is deployed more ubiquitously in public life, as well as in homes and schools. The President's Information Technology Advisory Committee recommended increased federal funding for research on the social and economic impacts of IT (PITAC, 1999). The NSF's Information Technology Research (ITR) initiative calls for research in several areas of IT that could be relevant to the problems identified in this chapter, including human-computer interfaces, information management, and the social and economic implications of IT.25 Its budget for the initiative was $90 million in FY00, to be used to support a mix of small (about $150,000 per year), medium (about $1 million per year), and large
($2 million to $4 million per year) projects. ITR could serve as a mechanism for supporting larger-scale efforts related to social applications of IT, perhaps even small centers. In addition, CISE expects to allocate $30 million to new information technology centers in FY00 to support fundamental research that spans the field of IT and encompasses scientific applications or addresses areas of social, ethical, and workforce issues (NSF, 2000). The challenge will be to find a suitable set of peers to review the proposals, drawing from the IT, social science, and domain-specific communities.
Interdisciplinary Research in Industry
Industry tends to be less wedded to a disciplinary research structure than universities are, and it has made a modest investment in interdisciplinary research motivated by the social applications of IT. Xerox Corporation, for example, has long kept social science researchers on the staff of its Palo Alto Research Center to help understand how people interact with IT systems in a variety of organizational settings. The work of the social scientists has been credited with improving the usability of Xerox copiers and streamlining internal processes for disseminating the knowledge of field service technicians (Bell et al., 1997). AT &T employs a number of economists to study the economics of the telecommunications industry, and several computer manufacturers, Apple Computer among them, have hired psychologists and cognitive scientists to inform their work on human-computer interfaces. In recent years, a number of Internet-based companies have recognized the need to make their Web sites more usable and have begun to hire employees with degrees in the social sciences and humanities.26 These efforts tend to focus on issues closely associated with product design and implementation and are diminutive in comparison to research on the purely technical aspects of IT.
As the market for IT-related services continues to grow, a number of traditional vendors of IT products are becoming more deeply entwined in the provision of large-scale IT systems and services. This trend could lead to greater investment in research on the social applications of IT. AT&T Labs, for example, recently announced a plan to fund research in the computer science and business departments of universities participating in the Internet 2 initiative, as a means of developing public key infrastructure for improving security across the Internet. The collaboration will be managed by a board of experts from industry, government, and academia who will invite participation from other players in those domains. In addition, MIT and Microsoft Corporation announced a partnership to develop educational technologies. Microsoft is investing $25 million in the venture over 5 years to pursue a range of projects, from online learn-
ing to new models for academic publishing. The projects will be managed by a steering committee consisting of members of Microsoft Research and MIT (Robinson and Guernsey, 1999).
IBM took such efforts a step further by establishing an Institute for Advanced Commerce within its Research Division in January 1998. The institute is a forum for examining fundamental shifts in business and trade in an information economy, with the goal of developing long-term, replicable electronic commerce solutions to meet corporate needs. It explores emerging economic trends and technologies to better understand the technical, business, and social processes that are shaping the electronic marketplace, and it has a business research center that studies the changing nature of work, industry structure, commerce, and technology. The institute began with an initial funding commitment of more than $10 million and is home to more than 50 researchers with expertise in IT, economics, and social science. Its board of directors consists of recognized IT researchers as well as IT executives from end-user industries such as automobile manufacturing and retail trade. Work is under way in areas such as (1) the evolving marketplace, (2) privacy, (3) variable prices and negotiated dealings, (4) managing the end customer, (5) the impact of globalization, (6) high-performance computing for commerce, and (7) system foundations (see Box 4.10 for a listing of recent projects in the institute). As of early 2000, the institute had produced more than a dozen reports on topics such as analyzing clickstream data to understand Web-based merchandising, business-to-business e-commerce, Internet auctions, and pricing in a free-market economy that contains software agents. It also has hosted several conferences on issues such as privacy in a networked world, gathering data on the information economy, and Internet-based negotiation technologies.
Industry efforts to conduct interdisciplinary research on the social applications of IT will undoubtedly be linked to business interests. Indeed, IBM is establishing an internal organization to help move new technologies into the marketplace by engaging customers in research. Of course, a key issue will be the ability of these programs to maintain a long-term outlook while producing more immediate results to suit ongoing business development opportunities. Time must pass before the program's effectiveness can be more rigorously evaluated.
Multidisciplinary Research Centers
The notion of interdisciplinary academic departments is realistic in the sense that similar organizations already exist, in the form of multidisciplinary research centers that involve technologists, social scientists, and end users. These centers range from centralized research facilities
Examples of Projects Pursued by IBM Corporation's Institute for Advanced Commerce
The IBM Corporation's Institute for Advanced Commerce has pursued projects in the following areas, among others:
SOURCE: IBM Corporation's Institute for Advanced Commerce home page at <http://www.ibm.com/iac/>.
that house researchers from many different disciplines under a single roof to virtual centers that enable interaction among researchers at different institutions. Their primary objective is to bring together researchers from the needed disciplines to jointly tackle a related set of problems, creating a critical mass of expertise to drive broad progress. A previous report by the National Academy of Sciences (1987) describes the benefits of the center mode of research as follows:
Centers contribute to science by enabling researchers to accomplish challenging, longer term projects that they could not undertake at all or as efficiently as individual investigators because of the need for stable support, large facilities or support teams, or simply the need to bring together diverse experiences and expertise. By involving external parties as well as students in their research activities, centers contribute to the more rapid transfer of new knowledge and to the training of professionals with an awareness of potential applications.
Experience to date suggests several factors that are important to the success of such centers: (1) capable and enthusiastic center directors, (2) high-quality investigators, (3) a mix of university and industry researchers, (4) a research agenda approved by all of the main participants, (5) acceptable arrangements to handle intellectual property and protect proprietary information, and (6) stable funding for long enough (perhaps 10 years) to produce real results.
In addition to those factors, multidisciplinary research centers for the social applications of IT might want to consider the following:
Focus—A center might need to focus on a particular application domain or a particular set of generic problems, such as e-commerce, medical informatics, or privacy.
Broad range of participants—To ensure that the centers conduct research motivated by social and economic needs as well as a desire for fundamental advances, they may need to link researchers from universities with industry vendors and experts from user organizations, including both companies and government agencies.
Links to testbeds or operational facilities—Each center would probably need to be linked to, and built on, large testbeds at one or more of the participants' sites so that researchers could have the sustained access to actual or near-operational systems or applications that is required to gain insight into real-world problems. The concern, of course, is that the individuals who manage major IT activities do not want researchers tinkering with their computers and software or taking up too much of their operators' time. One option is to give researchers access to these systems so they can observe but not alter them. Another option is to work with vendors and end-user organizations on new prototype systems and applications. A third option is to work on experimental testbeds that simulate real-world needs and operations.
Universities and industry have established a number of research centers to pursue multidisciplinary research on IT topics. The Information Networking Institute at Carnegie Mellon University, for example, was established in 1989 to examine the technologies, economics, and policies of global communication networks. The staff researchers have backgrounds in computer science, social science, cognitive science, and economics, and the institute has funding from both government and industry sources. The Media Lab at MIT, the Berkeley Wireless Research Center, and the Georgia Center for Advanced Telecommunications Technology (located at the Georgia Institute of Technology) also draw researchers from different academic disciplines and seek support from a number of industrial sponsors as well as from government. The main challenge for
such centers—in addition to the challenges of peer review and promotion outlined above—is funding. To support a staff of, say, 30 researchers plus graduate students, a center needs several million dollars a year, which tends to come from industry memberships or from project funding by either government or industry. Sustaining such funding can be a challenging task—especially if funding is sought from organizations (such as end users of IT) without a tradition of funding IT research.
At times, the government has helped establish research centers, either by concentrating research funding at particular universities (as DARPA has done with MIT, Carnegie Mellon, Stanford, and the University of California at Berkeley) or by establishing more formal programs that accept competitive proposals for the establishment or sustenance of a larger research center. The NSF, for example, has established programs to support science and technology centers (STCs) and engineering research centers (ERCs), both of which have been extensively evaluated and provide valuable lessons that can inform the establishment of other types of centers, such as those being considered under the NSF's ITR initiative (described above). The STC program is described below. Information on the ERCs is contained in Box 4.11.
The STC program, initiated in 1989, funds 28 centers that conduct interdisciplinary research in various fields of science (five of the centers conduct research directly applicable to IT).27 The STCs receive, on average, $2.3 million a year from NSF, plus funds from industry, the universities that host the centers, and other federal agencies.28 External reviews of the STC program have been favorable and have recommended continuation of the program. An assessment by the Committee on Science, Engineering, and Public Policy (COSEPUP) found that “most of the centers have been conducting outstanding research” and that “the STCs as a whole have done an excellent job of disseminating their results whether they are applied to basic science . . . or more applied fields” (NAS-NAE-IOM, 1996).29 Others have noted that the center mode of research is necessary to conduct large-scale, complex, interdisciplinary research such as that of an STC, and that the universities hosting STCs are removing traditional barriers between academic disciplines and are combating the biases against interdisciplinary work.
Some concerns have been raised about such centers: primarily the amount of time and energy dedicated to nonresearch missions. The NSF requires the STCs to engage in outreach activities, such as educational programs for grades K-12 in the communities in which they are located. The scale and type of the outreach programs vary from center to center, but their existence is a requirement for consideration for NSF funding (NSF, 1998). The COSEPUP report expressed concern that too much emphasis was being placed on community outreach and K-12 education at
The Engineering Research Center Program
The National Science Foundation (NSF) began funding engineering research centers (ERCs) in 1985 to create stronger links between industry and academic engineering programs and improve both the competitiveness of U.S. manufacturing industries and the quality of engineering education by making undergraduate and graduate training more relevant to industry needs. Thirty-four ERCs have been established. Each receives funding from NSF for 5 years, after which the funding can be renewed for another 5 years.1 Industrial participation is required. Companies that participate in ERCs must do more than contribute money: They must contribute intellectually, as well, to encourage interaction between students and representatives of industry. The ERCs average more than 30 industrial partners apiece.
Both an internal NSF review of the ERC program in 1997 and an external review by the National Academy of Engineering (NAE) in 1989 reported that ERCs contributed significant benefits to the nation, both economically and educationally.2 The NAE review concluded that the ERCs were responsible for novel research that was fundamentally important, making previously impossible interdisciplinary work feasible and providing experiences for students “that clearly excite them.” The NSF assessment reported that the ERCs discover new industry-relevant knowledge at the intersections of the traditional disciplines and transfer that knowledge to industry, while preparing a new generation of engineering leaders capable of leading in industry by engaging successfully in team-based, cross-disciplinary engineering to advance technology.
More than half of the firms involved in the ERCs that responded to an NSF survey reported that participation influenced the firm's research agenda, and two-thirds reported that ERC participation increased the firm's competitiveness. A majority of firms were able to improve products or processes, and 25 percent were able to create new products or processes as a result of ERC research. In addition, firms that employed graduates of the ERCs reported that the employees were “more productive and effective engineers than peers in the same firms. ” More than 80 percent of ERC graduates' workforce supervisors reported that the graduates were more prepared overall than their peers, contributed more technically, demonstrated a deeper technical understanding, were better at working in interdisciplinary teams, and had a broader technical understanding.
The NAE review questioned several aspects of NSF management of the program. Primarily, it was concerned about the fact that the NSF had chosen to reduce promised funding levels at ERCs to reduce costs per center. The review also reported that the ERC application process was too time consuming and that the selection process had been inconsistent over the years. Concerns over community outreach were not a problem because ERCs are not expected to engage extensively in such activities.
1The NSF funding averages about $3 million per center, which represents approximately one-third of the total funding for the centers, the balance coming largely from industry.
2The NAE report concludes with the following comment: “If the federal government is to assist industry in its fight to remain competitive, this is precisely the kind of program that it should support. If universities are to help build a technology base on which industry can draw, this is precisely the kind of role that they should play. And if industry is to take a hand in shaping policies that influence its long-term well-being, then here is precisely the way to become engaged.” See NAE (1989).
SOURCES: National Academy of Engineering (1989), National ScienceFoundation (1997).
the expense of research and recommended that the outreach continue but be given a lower priority (NAS-NAE-IOM, 1996). 30 Although they may spend considerable time on center activities such as these, the scientists involved report that they and their work benefit from the collegial interactions and exposure to the community and industry. 31 Also noted by COSEPUP was the importance of leadership to the success of an STC. If the center was to be more effective than just a group of individual researchers who happened to share a building and some equipment, its projects needed to be well integrated and effectively managed. Collaboration is an integral part of the center mode of research, but it must be nurtured, because most academic researchers are accustomed to working individually or in small teams.32 Lack of suitable leadership can undermine the value of a center.
Embedding Information Technology Research in Other Disciplines
As computing and communications have become embedded in many social applications, the role of computing in some disciplines other than computer science and engineering has changed and expanded. Because researchers in other disciplines are faced with designing systems in which IT is an essential element, they need to understand basic IT in its modern form—software applications distributed over a heterogeneous computer and network infrastructure—much better than is typically the case. They need to appreciate both the opportunities and the limitations of IT. An important responsibility will be the conceptualization and analysis of distributed information system requirements and specifications and the acquisition of sophisticated software applications through internal development, outsourcing, or purchase.
Because IT is becoming such a fundamental and pervasive aspect of many fields, it is natural for many departments on a university campus to become involved in research on the application of IT to their respective fields. It is becoming increasingly common for faculty and students in these departments to be facile with IT, and not infrequently the departments hire faculty members with a computer science background or degrees who have experience in the appropriate application areas. This is not a new phenomenon—it has a long history in other core disciplines such as mathematics and economics.33 Initially, collaboration with outside experts is a sufficient solution, but eventually the subject becomes important enough to deserve in-house expertise.
Accordingly, it can be expected that research related to the applications or implications of IT will be expanded in departments as diverse as engineering (and the subspecialties thereof), business, the social sciences, arts and performance, music, and others. One or more new disciplines
may arise out of these opportunities, much as computer science itself arose out of the collaborative efforts of mathematicians, electrical engineers, and physicists. This is a natural process that needs to be encouraged. Most likely some portion of the additional IT research funding being made available under the IT2 initiative will be devoted to cross-disciplinary research of this kind. Aside from setting up new programs (as outlined above), a number of other initiatives could help to establish interdisciplinary programs or to embed IT issues more firmly within other disciplines. For example,
Joint degree programs could be established between computer science and engineering and other relevant disciplines, such as the sub-specialties of engineering or business.
Restrictions on graduate programs in computer science could be relaxed to encourage students with backgrounds in other disciplines to pursue degrees. Conversely, students with undergraduate degrees in computer science and engineering could be recruited into the graduate programs of other disciplines.
Just as they have added expertise in mathematics and economics to other disciplines, universities could hire faculty members with strong backgrounds in computing for other departments, in part so that they could develop discipline-specific courses and teaching materials in the application of IT. Initially, many of these faculty members would probably have computer science degrees and work experience in a particular application domain; typical combinations might include business and transportation IT, computing embedded in mechanical systems, and so on.
Minors could be established in computer science and engineering programs and made available to students whose primary interests lie in other relevant disciplines.
Computing courses could be established for the benefit of a broad cross-section of students again modeled after mathematics and economics. Such courses could provide a breadth of understanding not available from more specialized courses.
Postdoctoral training programs could be established for social science or computer science Ph.D.s who wish to develop skills in research on IT in context.
In addition to promoting additional research on social applications, the expansion of interdisciplinary programs would help to redress the teaching and disciplinary imbalances that are likely to be created by increased student interest in IT. At the same time, this expansion would provide a badly needed influx of graduates with strong backgrounds in IT combined with domain-specific understanding. Many computer sci-
ence departments are today experiencing an enrollment surge similar to that of the early 1980s. Given current trends in technology, the wealth of job opportunities, and the excitement surrounding IT, this enrollment surge may be more permanent this time. Unfortunately, such growth often comes at the expense of other engineering disciplines, even though many of those disciplines continue to be vibrant and challenging and offer excellent job opportunities.
EXPANDING THE SCOPE OF INFORMATION TECHNOLOGY RESEARCH
The research programs described in this chapter are initial forays into the social applications of IT. In the process, many different models are being created that will coexist and complement one another. These research efforts will have to be redoubled, always informed by an awareness that multiple, complementary models of research exist. Researchers in traditional IT must broaden their outlook to encompass the social context for the technology, thereby changing what is thought of as the core of IT research.34 Conversely, researchers in other academic disciplines and end users of IT systems must become more actively engaged in IT research. To achieve the desired balance, new models for funding and conducting research must be explored. Only in this way will IT's potential to serve society be fully tapped.
Abt Associates, Inc. 1996 An Evaluation of the NSF Science and Technology Centers (STC) Program, Vol. I. Abt Associates, Cambridge, Mass.
Association for Computing Machinery (ACM), Special Interest Group on Human-Computer Interaction. 1992. Curricula for Human-Computer Interaction. ACM Press. Available online at <http://www.acm.org/sigchi/cdg/cdg2.html>.
Bass, L., P. Clements, and R. Kazman. 1998. Software Architecture in Practice. Addison-Wesley, Reading, Mass.
Bell, D.G., D.G. Bobrow, O. Raiman, M.H. Shirley. 1997. “Dynamic Documents and Situated Processes: Building on Local Knowledge in Field Service,” pp. 261-276 in Information and Process Integration in Enterprises: Rethinking Documents. T. Wakayama, S. Kannapan, C.M. Khoong, S. Navathe, and J. Yates, eds. Kluwer Academic Publishers, Norwell, Mass.
Benjamin, Matthew. 1999. “Clunky Internet Sites Get Social Science Treatment,” Investor's Business Daily, October 15, p. A6.
Bennis, W., and P. Ward Biederman. 1998. “None of Us Is As Smart As All of Us,” IEEE Computer 31(3):116-117.
Bernstein, P.A., and E. Newcomer. 1997. Principles of Transaction Processing. Morgan Kaufmann, San Francisco.
Bijker, W.E., T.P. Hughes, and T. Pinch, eds. 1987. The Social Construction of Technological Systems. MIT Press, Cambridge, Mass.
Brown, John Seely, and Paul Duguid. 2000. The Social Life of Information. Harvard Business School Press, Boston.
Cairncross, F. 1997. The Death of Distance: How the Communications Revolution Will Change Our Lives. Harvard Business School Press, Boston.
Clement, A. 1994. “Computing at Work: Empowering Action by Low-Level Users,” Communications of the ACM 37:52-65.
Cohen, W.M., and D.A. Levinthal. 1990. “Absorptive Capacity: A New Perspective on Learning and Innovation, ”Administrative Science Quarterly 35:128-152,.
Computer Science and Technology Board, National Research Council. 1989.Scaling Up: A Research Agenda for Software Engineering. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1992. Computing the Future: A Broader Agenda for Computer Science and Engineering. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1994. Information Technology in the Service Society: A Twenty-First Century Lever. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1996. Computing and Communications in the Extreme. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1997. More Than Screen Deep: Toward Every-Citizen Interfaces to the Nation 's Information Infrastructure. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1998a. Fostering Research on the Economic and Social Impacts of Information Technology: Report of a Workshop. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1998b. “Advancing the Public Interest Through Knowledge and Distributed Intelligence, ” Computer Science and Telecommunications Board, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 1999. Summary of a Workshop on Information Technology Research for Crisis Management. National Academy Press, Washington, D.C.
Computer Science and Telecommunications Board (CSTB), National Research Council. 2000. The Digital Dilemma: Intellectual Property in the Information Age. National Academy Press, Washington, D.C.
Davenport, T.H. 1993. Process Innovation: Reengineering Worth Through Information Technology. Harvard Business School Press, Boston.
Deck, Stewart. 1999. “Human Side Key to Data Warehousing,” Computerworld (February 15):14.
Diffie, W., and S. Landau. 1997. Privacy on the Line: The Politics of Wiretapping and Encryption. MIT Press, Cambridge, Mass.
Gibbs, W.W. 1994. “Software's Chronic Crisis,” Scientific American (September):86-95.
Grenier, R., and G. Metes. 1996. Going Virtual: Moving Your Organization into the 21st Century. Prentice-Hall, Englewood Cliffs, N.J.
Grief, Irene, ed. 1988. Computer-Supported Cooperative Work: A Book of Readings. Morgan Kaufman, San Mateo, Calif.
Hutchins, E. 1991. “Organizing Work by Adaptation,” Organization Science 2:14-39.
Iacono, S., and R. Kling. 1984. “Office Routine: The Automated Pink Collar,” IEEE Spectrum(June):73-76.
Jones, C. 1996. Applied Software Measurement. McGraw-Hill, New York.
Keen, P.G.W., and C. Balance. 1997. On-Line Profits: A Manager's Guide to Electronic Commerce. Harvard Business School Press, Boston.
Kleinrock, L. 1974. “Research Areas in Computer Communication,” Computer Communication Review 25(1):33-35, July.
Kling, R., and W. Scacchi. 1984. “The Web of Computing: Computer Technology as Social Organization, ” Advances in Computers 21:1-90.
Markus, M.L. 1987. “Toward a ‘Critical Mass' Theory of Interactive Media: Universal Access, Interdependence, and Diffusion,” Communication Research 14:491-511.
McKnight, L.W., and J.P. Bailey. 1997a. “Internet Economics: When Constituencies Collide in Cyberspace,” IEEE Internet Computing 1(6):30-37, December.
McKnight, L.W., and J.P. Bailey. 1997b. Internet Economics. MIT Press, Cambridge, Mass.
Messerschmitt, David G. 1999. Networked Applications: A Guide to the New Computing Infrastructure. Morgan Kaufmann Publishers, San Francisco.
Mitchell, W.J. 1996. City of Bits: Space, Place, and the Infobahn. MIT Press, Cambridge, Mass.
National Academy of Engineering (NAE). 1989. Assessment of the National Science Foundation's Engineering Research Centers Program. National Academy Press, Washington, D.C.
National Academy of Public Administration (NAPA). 1995. National Science Foundation's Science and Technology Centers: Building an Interdisciplinary Research Paradigm. NAPA, Washington, D.C.
National Academy of Sciences (NAS), Panel on Science and Technology Centers. 1987. Science and Technology Centers: Principles and Guidelines. National Academy Press, Washington, D.C.
National Academy of Sciences-National Academy of Engineering-Institute of Medicine (NAS-NAE-IOM), Committee on Science, Engineering, and Public Policy. 1996. An Assessment of the National Science Foundation's Science and Technology Centers Program. National Academy Press, Washington, D.C.
National Science Foundation (NSF). 1997. The Engineering Research Centers (ERC) Program: An Assessment of Benefits and Outcomes. NSF, Arlington, Va.
National Science Foundation (NSF). 1998. Science and Technology Centers (STC): Integrative Partnerships. Program Solicitation NSF 98-13. Available online at <http://www.nsf.gov/od/oia/stc/nsf9813.html>.
National Science Foundation (NSF). 1999a. Digital Government. Program Announcement NSF 99-102, Computer and Information Science and Engineering Directorate. Arlington, Va.
National Science Foundation (NSF). 1999b. “National Science Board Approves Five New NSF Science and Technology Centers,” Press Release, July 29. Available online at <http://www.nsf.gov/od/1pa/news/press/99/pr9945.htm>.
National Science Foundation (NSF). 2000. FY2001 Budget Request to Congress. Available online at <http://www.nsf.gov/home/budget/start.htm>.
O'Leary, D.E. 1998. “Enterprise Knowledge Management,” IEEE Computer 31(3):54-61.
Orr, J. 1990. “Sharing Knowledge, Celebrating Identity: War Stories and Community Memory in a Service Culture,” pp. 169-189 in Collective Remembering: Memory in Society, D.S. Middleton and D. Edwards, eds. Sage Publications, Beverly Hills, Calif.
President's Information Technology Advisory Committee (PITAC). 1999. Information Technology Research: Investing in Our Future. National Coordination Office for Computing, Information, and Communications , Arlington, Va., February. Available online at <http://www.ccic.gov/ac/report/>.
Resnick, P., and H. Varian, eds. 1997. “Special Section: Recommender Systems,” Communications of the ACM 40(3):56-58.
Robinson, Sara, and Lisa Guernsey. 1999. “Microsoft and MIT to Develop Technologies Together,” New York Times Cybertimes, October 5. Available online at <http://www.nytimes.com/library/tech/99/10/biztech/articles/05soft.html >.
Rosenberg, R.S. 1997. The Social Impact of Computers, 2nd ed. Academic Press, New York.
Schofield, J. W. 1995. Computers, Classroom Culture, and Change. Cambridge University Press, Cambridge, United Kingdom.
Shapiro, C., and H. Varian. 1998. Information Rules: A Strategic Guide to the Network Economy. Harvard Business School Press, Cambridge, Mass.
Siegel, J.L., V. Dubrovsky, S. Kiesler, and T. McGuire. 1986. “Group Processes in Computer-Mediated Communication,” Organizational Behavior and Human Decision Processes 37:157-187.
Sproull, L.S., and S. Kiesler. 1991. Connections: New Ways of Working in the Networked Organization. MIT Press, Cambridge, Mass.
Stonebraker, Michael. 1998. “Are We Working on the Right Problems?” SIGMOD Record (ACM Special Interest Group on Management of Data) 27(2):496.
Stonebraker, M., P.M. Aoki, R. Devine, W. Litwin, and M. Olson. 1994. “Mariposa: A New Architecture for Distributed Data,” pp. 54-67 in Proceedings of the 10th International Conference on Data Engineering, IEEE Computer Society Press, New York.
Wald, Matthew L. 2000. “New Systems for Controllers May Ease Air Traffic Woes,” New York Times, March 23. Available online at <http://www.nytimes.com/library/tech/biztech/articles/23traffic.html >.
Walton, R.E. 1989. Up and Running: Integrating Technology and the Organization. Harvard Business School Press, Cambridge, Mass.
Washington Technology. 2000. “Federal Agency IT Expenditures,” January 10, p. 6.
Wulf, William. 1999. “A Call for Technological Literacy,” IEEE Computer Applications in Power 12(2):10.
1. The President's Information Technology Advisory Committee (1999) identified nine major transformations that IT will bring to society.
2. Other examples can be found to further demonstrate the introduction of IT into an existing application. One example is remote conferencing, which is intended to reproduce and improve on face-to-face meetings or voice-only teleconferencing as a means for group interaction. In business, enterprise resource planning applications are intended to improve standard business processes such as human resources, finance, and sales, building on a history of more focused multifaceted systems for manufacturing resource planning.
3. This topic was the subject of an earlier CSTB study. See Chapter 4 of CSTB (1994).
4. For example, the acquisition and use of word-processing applications in organizations are affected by status hierarchies. At one point managers bought word-processing systems that were used by the word-processing pool. The users, lower status clerical personnel, had no control over what was used or its conditions of use (Iacono and Kling, 1984). Suggestions for improvements in the application or its conditions of use that were made by users were ignored because the users had low status (Clement, 1994). As the function of word-processing clerical personnel was taken over by white collar workers doing their own word processing, those white-collar workers encountered hidden interdependencies. For example, people could not exchange documents with others who were using incompatible software. Sociologists have been conducting analyses of the socially embedded nature of (apparently) stand-alone systems since the early 1980s (e.g., Kling and Scacchi, 1984). Also see Bijker et al. (1987) for a more general example of the social construction of technological systems.
5. The challenges and opportunities for designing systems that support a larger and
diverse population of users in a larger variety of applications, focusing on issues of usability, are outlined in CSTB (1997).
6. It is fairly easy to understand the capabilities needed by a word processing program or spreadsheet, although as collaborative authoring features have been added, these capabilities have become increasingly sophisticated and complex. A scientific computation has a relatively well-defined form and capability. Even a business application such as transaction support or management of accounts receivables or payroll records are reasonably well defined and understood from the outset. Broader application concepts, epitomized by the expansive term “ e-commerce,” inherently embrace numerous interactions among systems, organizations, and individuals at multiple levels.
7. As used here, the term “economic” refers to a broad range of purposeful activity, including not only activity associated with various goods- and services-producing industries but also that associated with research, learning, and government.
8. “Knowledge management” is a new term that has the disadvantage of being a management buzzword, with the attendant hype, but if the hype can be set aside, the concept can be leveraged to set ambitious objectives for making better use of information through technology.
9. See Stonebraker (1998) and Stonebraker et al. (1994).
10. For a detailed discussion of the complexities of intellectual property protection in a digital environment, see CSTB (2000).
11. The core role of governments is affected by IT, particularly by the global nature of networks. The concept of sovereignty rests largely on the geographical separation of jurisdictions, which is undermined on a global network. The trend toward more international governance mechanisms to deal with international issues is a natural response to globalization, but the trend is accelerated by global computer networking and the applications it supports. Issues such as privacy, restricted access for children, and taxation demand not only new governance mechanisms but also new technologies to support them.
12. Here again, networking is a driver, building on historical improvements in transportation and telecommunications and resulting shifts in markets, organizational scope and scale, and institutional relationships. Business processes and relationships associated with contemporary IT cannot be appreciated accurately without acknowledging that history.
13. The CSTB developed a powerful illustration of the value of systematic study of a specific application domain. Asked to look at crisis management, the board intrigued a group of computer scientists with no knowledge of that domain by exposing them to the problems of people whose jobs revolve around planning for, and responding to, civilian and military crises (CSTB, 1996). The communication about the problems inherent in crisis management, in turn, led to new computer science research. Some of the problems and solutions were common to those found elsewhere, but even some of those had domain-specific requirements, as evidenced by a project participant's observation that some of crisis management technology was like “digital libraries with deadlines.”
14. It must be emphasized that social applications research is about technology as well as social, economic, and political systems. Its goal is to make technological progress more dependent on visionary attention to the uses and needs for that technology and not simply on a near-term, incremental commercial and technical research agenda. As IT is encumbered with few fundamental limits and is mostly what we make of it, the goal is to aim technological advances in directions that offer the most benefit to society. This research is not only about the impact of technology on society, as emphasized by the report of the President's Information Technology Advisory Committee (1999), but also about the impact of society and humanity on the requirements of future technologies, with the aim of maximizing the beneficial impact and minimizing the harmful ones.
15. These figures are from the Office of Management and Budget as cited in Washington Technology (2000).
16. The Digital Government program is administered by the Computer and Information Sciences and Engineering (CISE) directorate at NSF but grew out of an effort by the Federal Information Services and Applications Council (FISAC) of the Computing, Information, and Communications Research and Development (CIC R&D) Subcommittee of the National Science and Technology Council. FISAC was created to stimulate and foster the migration of technology from the IT community to government application missions and information services communities and to identify challenges from applications to the IT R&D community. It has participants from across the federal government (including the Department of Health and Human Services, the Department of Defense, the National Aeronautics and Space Administration, the Department of Agriculture, the Department of Transportation, and the Environmental Protection Agency). It (1) promotes the early application of advanced computing, information, and communications technologies and R&D capabilities to critical federal government missions, (2) supports multiagency leadership in efforts that demonstrate, deploy, and implement advanced computer and information technologies that have the potential to be widely applicable to federal agency missions, (3) encourages pilot projects to assess the critical computing, information, and communications technologies (e.g., security technologies) needed by applications, and (4) supports broad administration goals in the international arena that eliminate barriers to applications. It drew inspiration from CSTB (1996).
17. Additional information on requirements for the Digital Government program is available in the program solicitation. See NSF (1999a).
18. For example, the first solicitation, in September 1998, attracted 50 proposals (many of them for planning grants) that involved, among others, the Bureau of the Census, the Bureau of Labor Statistics, the Federal Emergency Management Agency, the Coast Guard, the National Cancer Institute, the Department of Justice, the National Oceanographic and Atmospheric Administration, the U.S. Geological Survey, the Department of Energy, the Department of Housing and Urban Development, the General Services Administration, the Federal Reserve Bank, the National Institute of Standards and Technology, the National Security Agency, the Office of Management and Budget, and the Environmental Protection Agency.
19. Information on IBM's FOAK program was provided by Armando Garcia, IBM Corporation, personal communication dated July 28, 1998, and by Carol Kovak, IBM Corporation, personal communication dated April 20, 2000.
20. As an example of this phenomenon, consider the case of the Computers, Organizations, Policy, and Society (CORPS) group within the Department of Information and Computer Science at the University of California at Irvine. CORPS concerns itself with studies of the organizational, economic, and social aspects of computing and has strengths in human-computer interaction, computer-supported cooperative work, and information retrieval. When the department was reviewed in 1997 as part of a mandatory 5-year external review of its research and graduate programs, the review committee (which consisted primarily of respected computer scientists) recommended that CORPS be removed from the department and placed somewhere else in the university, not because the research was weak (on the contrary, the review committee declared it to be excellent and important), but because the researchers used perspectives informed by the social sciences and therefore could not understand the engineering perspective at the heart of computer science. The department did not take the review committee's advice (on this subject at least), but the case demonstrates the challenges of rewarding interdisciplinary efforts in the framework of highly specialized disciplines.
21. These and other attitudes and perceptions about barriers to cross-disciplinary collaboration were elicited by CSTB Director Marjory Blumenthal through conversations with
MIT faculty members in computer science, electrical engineering, and social sciences, as well as administrators, in 1998.
22. Historically, most civil engineering departments have been, by necessity, sociotechnical systems departments. Similar divisions have been tried at many universities. In the 1960s and 1970s, the Sloan Foundation funded a number of universities to create divisions of this type. Other examples include MIT's Energy Laboratory and Carnegie Mellon University's Robotics Institute, which go back 25 years or more. Historically, these centers flourished for a number of years and then either atrophied or faded as funding shifted to different sociotechnical system areas or the faculty champions retired.
23. The NSF has also supported some work on sociotechnical systems through its Directorate on Social, Behavioral and Economic Sciences (SBE), but most of that work has focused on issues not directly associated with IT. Some researchers at the nexus of computing and the social sciences claim that SBE has not been supportive enough of the centrality of technology in such research, but in recent years, the directorate has cosponsored work (with the CSS program) on research challenges related to the social and economic impacts of IT on intellectual property protection in a digital environment. See CSTB (1998a) and CSTB (2000).
24. A “Dear Colleague” letter posted on the CISE Web site in 1999 noted that increased future funding was anticipated for the CSS program and called for proposals related to traditional CSS interests and the broader issues of social and economic implications of IT. Proposals could request up to $300,000 in funding for 36 months. The CSS expected to make about 10 awards in FY99. See National Science Foundation, Computing and Information Science and Engineering Directorate. Undated. “Dear Colleague” letter from Michael Lesk, division director, Information and Intelligent Systems Division. Available online at <http://www.interact.nsf.gov/cise/html.nsf/html/css_dcl?OpenDocument >.
25. Indeed, the purpose of NSF's ITR program is to “enhance the value of information technology for everyone.” The complete list of areas in which NSF is soliciting proposals under the ITR program is as follows: software, IT education and workforce, human-computer interface, information management, advanced computational science, scalable information infrastructure, social and economic implications of IT, and revolutionary computing. Letters of intent for proposals exceeding $500,000 were due in November 1999; those for smaller projects were not due until January 2000. The NSF anticipated making awards under the ITR program in September 2000. See NSF (1999a).
26. Many Web designers do not understand user behavior, including why users often leave sites soon after going to them. In an attempt to understand a user's experience of a Web site, Modem Media uses the technique of role playing, in which employees pretend to be users that fit a certain profile. Modem Media intends to hire psychologists and anthropologists to expand its efforts to understand user behavior. Meanwhile, Sapient announced plans to buy E-Lab because of E-Lab 's knowledge of “patterns of behavior that reveal and drive the nature of experience. ” However, Web site usability expert Jakob Nielsen says social scientists are not the answer, and that companies should focus instead on conducting usability tests with actual customers. See Benjamin (1999).
27. The STC program was initiated in response to President Ronald Reagan's 1987 State of the Union address, which proposed the establishment of federal centers to promote U.S. economic competitiveness. Of the original 25 centers funded from the first two program solicitations, 23 remain; 5 new centers were granted funding in July 1999. See NSF (1999b).
28. Industrial support is not a requirement for the centers, but the STCs average eight industrial partners per center. See National Academy of Public Administration (1995).
29. A bibliometric analysis conducted by Abt Associates, Inc., found that the journal publications of STC researchers were cited 1.69 times as often as the average U.S. academic paper published and that the journals in which STC scientists published had greater influence than the average scientific journal. In addition, papers from STC researchers are cited
two to four times as often in U.S. patents as the average academic research paper. See Abt Associates (1996).
30. In most cases, outreach did not significantly interfere with the conduct of research, but it did in the case of at least one STC. The center was told after one NSF site visit that it did not conduct enough outreach programs, so it began participating in so many outreach activities that after another NSF site visit, the center was told that it did not conduct enough research. In response to the first site-visit report, almost all of the time of the postdoctoral scientists and almost all the resources of the center in the summer months were devoted to K-12 outreach programs.
31. Most of the time is spent by a center's director and administration, and only a small proportion of the time is spent by the actual researchers.
32. The Abt Associates evaluation of the STC program was even more favorable than the COSEPUP report. Abt writes that “individual centers have produced significant research achievements in fundamental knowledge and the development of research tools, and have identified a range of downstream impacts of this work.” Abt found the centers to be particularly flexible and effective in responding to scientific opportunities and reported that “industry partners consider their affiliations with the STCs to be immensely beneficial.” In addition, the Abt report viewed the K-12 educational programs more favorably than did the COSEPUP report. See Abt Associates (1996).
33. As an analogy, consider the roles of mathematics and economics in other fields. As they became increasingly critical to a number of fields—mathematics to physics and economics to business or agriculture, for example—collaboration as a way of addressing the resulting challenges soon became inadequate. Rather, domain experts felt the need to become sufficiently adept at mathematics or economics to contribute directly in these areas. The situation is similar with IT, which is becoming an integral part of the sociotechnical applications within which it is embedded; that is, the artificial separation of application-specific and information technology expertise is no longer effective. A solid grounding in mathematics is considered essential to the natural sciences and engineering, and a solid grounding in economics is considered essential to business, agriculture, and a number of other fields. Similarly, modern forms of IT should be considered normal and essential parts of the background in a number of other fields. These fields include particularly the education of future engineers in fields as diverse as civil, mechanical, aerospace, electrical, and nuclear engineering, and also the education of future business managers (many of whom specialize in the social sciences and humanities as well as business). A broad cross-section of students in other natural and social science disciplines also need to take at least a foundation course in these technologies, analogous to a first course in economics.
34. This continual expansion of what is considered the core is healthy and needs to be strongly encouraged. Two fields that were once considered applications of computers are computer graphics and database storage systems. After computer science researchers began to make progress on these topics and publish papers and the capabilities became a normal part of many systems, they gradually came to be considered part of the core of the computer science research community. The technologies surrounding e-commerce are in the process of making this transition. Controversy surrounding the publication of the CSTB report Computing the Future: A Broader Agenda for Computer Science and Engineering in 1992 suggests that change is not always welcome or even understood. See CSTB (1992).