Participants emphasized that "There are other knowledge networks besides the Internet." Knowledge networks based on non-electronic technology have long existed, and may provide valuable case studies for gaining insight into the nature of knowledge networks developing under electronic technology.

1.1 LEARNING FROM EXISTING OR PRIOR EFFORTS AT KNOWLEDGE NETWORKING

Recommendations:

Systematic data collection from existing knowledge networks based on electronic technology, including comparative analysis of case studies, in order to glean insights for policymaking, research direction, planning, and organization of NSF knowledge networking initiatives and for the design of knowledge networks themselves.

Comment:

Obviously, no current or previous knowledge network quite parallels the type of broadscale knowledge network that the NSF initiative envisions. Workshop participants, however, appeared to believe strongly that NSF should avoid "reinventing the wheel" or repeating the mistakes of the past, and that systematic collection of data from similar knowledge networks (including comparative case studies) would be of considerable help in planning and development. They appeared to be suggesting research that is short-term, integrated closely with the planning for the knowledge network, and directed toward identifying specific parameters, conditions, etc., for success or failure (which also requires defining "success" and "failure" for knowledge networks). [Examples mentioned included: computational chemistry's CCL (8000 participants); high energy physics' HEPnet preprints; Ernst and Young's Knowledge Base; Eureka (repository vs. tandem tech-list); Pueblo K-6 Longview School; Amazon.com (book readers and buyers); Health Support Usenet groups; community nets.

Some of the participants emphasized that data collection should extend to examples from beyond the academic scientific community because the knowledge networks envisioned will themselves stretch well beyond the scientific community and planning and organization should reflect more than academic scientists' concerns and work patterns.

Examples of research suggested:

Comparative, intense studies of knowledge networks, with the goal of identifying "those principles of socio-technical design that are robust now" and from which we might discern patterns of development.

Study examples of where the scientific community has reacted to a short-term situation or a new phenomenon (e.g., a volcanic eruption, astronomical event, or epidemic) by establishing a "knowledge network". What worked well? (Some such histories exist.)

Integrate historical data for hypothesis testing and research on the "here and now," for analysis to detect patterns, or for identifying testable questions related to predictability of success or failure.

How might "disparate data bases" of existing case studies be integrated and organized?

What is the nature of scientific collaborations? Can one identify the necessary and sufficient conditions that determine the success of scientific collaborations and in particular, what makes collaborations fail? Study of failures or anomalies in similar knowledge networks. Choose some interdisciplinary projects that worked (or didn't work) and disaggregate them, "reverse engineer" them.

How have users differed in their participation in knowledge network? by status and role? by cognitive style? by type of interest?

How is information shared in existing knowledge networks -- and with what consequences?

Compare how participation has changed (increased or decreased) in various knowledge networks. What factors affected that change, other than new information technologies?

What is the correlation between flexible organization and longevity of knowledge nets?

To what extent is knowledge embedded, or readily available, in successful knowledge networks?

In knowledge networks generally, What are the communications paths? What type of knowledge do they propagate? What are their sizes? How much do they rely on volunteerism? How porous or permeable are their borders? How flexible are their organizations? What historical-sociological frameworks (models) might be developed for knowledge networking?

Are there are appropriate theoretical frameworks, based on concepts of computation, information, and knowledge, that are appropriate for modeling knowledge networking? For example, is the emerging computational theory of knowledge an appropriate basis?

Recommendations:

NSF should pay close attention to social science data that already exists (e.g., about the nature of scientific communication, or about the nature of intellectual or research collaborations) and could be synthesized and applied in planning for knowledge networks or could be tested as knowledge networks are developed.

Comment:

This was a clear plea to NSF that planning, policymaking, and research direction include careful consideration of what we may already know about various aspects of knowledge networking from decades of social science research. We note that much relevant communications and research -- especially that relating to mass and electronic communication -- has been funded by sources other than NSF, such as the communications industry and private foundations. In addition, communications research generally has tended to ignore science content and so the number of science communications researchers and scholars has been relatively small. This latter situation has been changing in the past decade and a body of quantitative research is being developed.

Examples of research suggested:

Review and synthesis of existing research, including identification of: -- where there is a need for more quantitative data, theory-testing, and -- where there is congruence of factors for success.

Development of protocols and testing (experimental methods, computer simulation ...) of data from these comparisons or from meta-analyses.

What data exist on how, when, and why collaborations succeed or fail?

What data exist on the conditions for broadscale, diverse communication among expert groups and users of expert knowledge?

What data exist on the conditions for successful interdisciplinary collaboration?

A critical set of issues that recurred throughout the workshop was the need to support and catalyze the construction of knowledge networking environments, for purposes of both research and application.

2.1 DEVELOPMENT OF APPROPRIATE RESEARCH METHODOLOGIES

Recommendations:

NSF should support research that is aimed at developing the appropriate research methodologies for studying the human dimensions of knowledge networking.

Comment:

There is a need for research related the special methodological problems arising in research on knowledge networks (e.g., need to study online behavior and therefore address special issues of privacy, anonymity and pseudonymity, etc.)

Examples of research suggested:

How does one determine the analysis units for knowledge networks?

How can one represent the relationships between nodes?

How may compiling extensive data on users' attitudes intrude upon individual rights to privacy or imply threats to their freedom of expression?

Recommendations:

NSF should support the development or designation of specific knowledge networks as testbeds for the study of social or psychological factors in collaborative knowledge networking, in addition to study of technical and design issues is a critical issue. NSF should allocate funding specifically for this purpose. In particular, research issues should relate both to the nature of such testbed systems and to the classes of issues that they may be employed in resolving.

Comment:

It is important that NSF support the implementation of electronically-based knowledge networking testbeds that significantly extend current digital library functionality to collaborative work environments. Such testbeds should support both the activities of significant communities in "real" applications as well as the activities of scientists investigating the nature and efficiency of such activities. At the least, NSF should consider providing or subsidizing tools that support the development of knowledge networking infrastructure.

Examples of research suggested:

What are the characteristics of a knowledge network that make it differ from the current Internet/Web environments?

Is it the fact that knowledge networking requires a semantic/interpretive layer about the "syntactic" layer? What tools/techniques should be used to create this semantic layer?

What are the characteristics that such a testbed should have in terms such characteristics as the heterogeneity of the nodes, the distribution in time and space of the nodes, the overlap of multiple and possibly hierarchical networks, the size of the network, the application areas that should be supported, etc.?

What metrics of performance should be established for such testbed systems?

Study processes of knowledge acquisition, organization, dissemination and use for emerging knowledge networks.

Can the introduction or use of new information technologies in knowledge networks lead to losses of knowing, and if so, how may this be measured?

In an ever-expanding knowledge networking context, with no natural endpoint to processes or system and with no single goal, what is the relationship between closure and coherence? Is closure necessary for coherence? How may lack of closure influence collaboration?

Are there potential downsides (hidden costs) to connectivity, to formalizing connections or to "computerizing" (de-personalizing or de-humanizing) them? For example, does connectivity increase knowledge dependency because people prematurely discard earlier technologies or earlier connections?

Recommendations:

Research is needed on the skills needed now, as well as on the skills likely to be needed, by the next generation of researchers into knowledge networks.

Comment:

We need a new generation of knowledge networking researchers to create and study the implications of knowledge networks, because such work requires new and flexible interdisciplinary and multidisciplinary approaches. These researchers will form part of the knowledge networking infrastructure.

Examples of research suggested:

What do knowledge network researchers need to be able to do? What institutional organizations/frameworks are best suited to train individuals with the appropriate sets of skills to perform valuable research in knowledge networking? Funding for scholarships and for graduate and post-graduate fellowships. Funding for faculty development in the teaching of knowledge networking skills.

How can professional knowledge networks be leveraged to apply in teaching/learning situations broadly? Creation of centers for knowledge networking training and education.

Recommendations:

Research is needed on the skills likely to be needed by the users of knowledge networks and on the environments that would help users to learn such skills.

Comment:

We need to study how current and future users of knowledge networks behave in such environments and how their use of the networks can be made effective and efficient.

Examples of research suggested:

How are people learning in (through) existing knowledge networks? Would this change as research network activities become larger and more complex?

What are appropriate institutional models for training, especially alternative models, and work that complements development of virtual universities?

What are best arrangements for creating value for knowledge networks and for capturing that value and hence providing best possible environment for knowledge networks?

At what level should people learn knowledge networking skills?

How should the nation balance the need for specialists with need for interdisciplinary researchers in the knowledge network context?

As new roles emerge within knowledge networks, will they require new skills? if so, what? how will skills be acquired? standardized?

Identify alternative institutional arrangements for interdisciplinary training, partnerships, curriculum development.

Identify purposes and goals of existing interdisciplinary partnerships and how they relate to design, success, etc.

How can we develop "literate consumer" for knowledge. Should such development be the responsibility of the provider or manager of a knowledge network? If so, how should such consumption skills be developed?

Discussions within the workshop breakout groups (see Appendix A) focused on a wide variety of issues, including policy, sociological, psychological, and philosophical issues, as well as technical considerations relating to the organization and functioning of knowledge networks. Our presentation of the discussion of these issues is ordered alphabetically.

Comment:

Some participants emphasized that "organizational decisions about knowledge networks need to be made with underlying values in mind because form affects function" and that a central set of issues will involve "how to make explicit (a) the implicit knowledge and (b) the values of artifacts and social-machine systems? " Some participants also noted that many of the issues discussed below, such as access, authority, governance, responsibility, and trust, may be viewed as social or political "values" that should be investigated from the normative viewpoint, as well as from the descriptive viewpoint which relates, for example, to how they affect the performance of a knowledge network.

3.1 ACCESS ("Access," one participant noted, "is not only [a matter of] downloading knowledge.")

Recommendations:

Research on parameters and conditions for access to participation in knowledge networks and access to the knowledge bases created within or by them.

Examples of research suggested:

Is a knowledge network a public or a private forum?

To what extent is access to knowledge network collaborations a factor of access to or acquisition of tools? of skills? or other factors?

What are the important issues of usability in providing access to the physically disabled? to members of low-income or socially disadvantaged minorities?

What is the relationship between the degree and (tightness) of access and the quality of the knowledge network?

What is the relationship between a knowledge network design and the extent to which it inhibits or facilitates participation by women?

What artificial barriers to full public participation might be inadvertently erected in knowledge networks, such as designs that require the most advanced browsers or searching software?

What is the relationship between access to knowledge networks, the knowledge in them, and the ability to build them?

Investigate issues of licensing and credentialing.

Who can use the knowledge? who can contribute to the base? whose knowledge or knowledge processes have authority in the system?

Importance of gatekeepers.

How is access linked to the establishment of credibility, or to establishment and determination of authority?

Recommendations:

Research on how authority is asserted, established, sustained, or dismissed in knowledge networks, and on the linkage between social authority and the technical parameters for assigning authority (e.g., precedence).

Examples of research suggested:

What are the models for the assertion and establishment of (social) authority in existing knowledge networks?

In knowledge networks, is the authority of knowledge hierarchical?

To what extent does the authority of knowledge mimic (follow) the technical model adopted for the computer system(s) used?

What are the conflicts between shared purpose and authority How is power (or control) negotiated in a knowledge network--and with what consequences?

How can alternative points of view be effectively heard in a knowledge network without endangering a spirit of community?

How do (can) knowledge networks accommodate multiple stakeholders? How is authority maintained independent of the network authority?

Recommendations:

Research on how creativity and innovation are stimulated and developed in interdisciplinary knowledge networks.

Examples of research suggested:

What factors influence creativity and innovation in a collaborative, interdisciplinary research environment?

How can creativity be fostered not just rewarded after it has reached application?

Recommendations:

Research on the social structure of knowledge networks and the relationships among networks sub-groups and sub-cultures.

Examples of research suggested:

What tools can help create the semantic layer? What norms of reciprocity may be needed for successful knowledge network?

How can these norms be established for the group?

What impact may large, complex and all-inclusive knowledge networks have on the survival of smaller knowledge networks?

To what extent is group coherence or identity related to community or group autonomy or independence (especially if knowledge networks are composed of coalitions of previously established small knowledge networks)?

How can small group connectivity be maintained in a large-scale knowledge network?

How can responsibility and accountability be embedded within organization of a knowledge network?

To what extent might data from ethnographic studies of tribal communities (or of scientists) be applied (or hypotheses tested) through study of knowledge network subgroups?

What are the similarities and differences between the current and the accepted norms or work standards for scientific research and the norms and standards for net culture? Where are points of predictable conflict?

How might knowledge networking affect those who are not connected? or those who have dropped out?

Recommendations:

Issues of credit and reward, incentives and disincentives, for interdisciplinary research must be taken into account as part of knowledge networking activities in academic environments.

There were two distinct parts to these recommendations: (a) NSF should, to the extent that any government agency can, attempt to insure that engaging in interdisciplinary knowledge networking research, does not place participants at a professional disadvantage, and (b) There is a need for research on the incentives and motivations, credits and rewards, for those who establish or participate in knowledge networking.

Examples of research related to recommendation (b):

What are the incentives structures to participate in knowledge networking, and how are they likely to be changing as the knowledge networking concept develops?"

"How will people who engage in interdisciplinary knowledge networking projects get credit and reward in 'realtime' organizations?" How will the best researchers not only "exist" but thrive in discipline-centric institutions and systems (e.g., prizes, fellowship competitions, and research grant review panels) while engaging in interdisciplinary work of the highest quality?

How can excellence in interdisciplinary research be identified? Are interdisciplinary researchers inevitably in a "double bind," that is, always judged by the highest standards of both (or all) possible disciplines yet dismissed by all because of the interdisciplinary nature of the questions?

Recommendations:

Research on how the interrelatedness ("interleaving") of work and leisure activities in electronic communication venues may influence aspects of knowledge networking, from skills acquisition to the participants' expectations of venues and content.

Comment:

Participants agreed that knowledge networking -- no matter how serious, important, and cerebral the activities and purpose might be -- will nevertheless take place in a computer networking domain that is already significantly influenced by the values, techniques, motivations, policies, etc., of entertainment and the search for leisure. There is already a substantial body of research on similar issues, much of it on television and video games, and directed at childhood socialization and education issues, and some from operator training research, which might be built upon.

This is a difficult, complex and important issue which now takes on serious import as the same communications channels are being utilized for work, play, education, and entertainment.

This is also an issue of importance for public communication of science and for civic discussion of science and development of science literacy.

Examples of research suggested:

Study how the growing interrelation of work and leisure activities on computer networks may affect the standards, expectations, norms, and values of scientific knowledge networks.

To what extent will knowledge networking skills and techniques be influenced by the ubiquity of entertainment-based or play-based activities in modern American life, especially on Internet and WWW?

How will education and training for knowledge networking be influenced by, for example, graphics designed for computer games?

What are the implications of this blurring of the boundaries between work and leisure, or of the interleaving of these types of experiences and expertise?

What about the leakage (or deliberate adoption) of standards from one domain to another?

Will (should) knowledge networks accommodate to the influence of entertainment by adopting or adapting standards from other domains (e.g., from entertainment)? Can they survive by adhering only to the standards of science and the academy?

What can we learn from the entertainment domain? If knowledge networks are evolving from a technical, work-oriented standard to one in which standards, tasks, roles, ethical principles, and intellectual property rights are influenced by the entertainment or leisure industries, then what role might anticipatory research play in shaping this evolution?

Recommendations:

Research on how organization and governance issues may differ when interdisciplinary activities take place in a networked environment.

Examples of research suggested:

What sort of organizational environments should we design for doing interdisciplinary work, for sharing ideas, for interdisciplinary activity?

How are existing knowledge networks organized and managed? Identify successful models for organization and management through historical case studies, testing and experiment.

Are there factors other than participants' skills, disciplinary affiliations, etc., which may affect success in international knowledge networks, such as cultural differences in how research is organized and funded, funding cycles and priorities, systems for credit and reward, attitudes toward intellectual property, gender roles, or freedom of expression?

What are the implications of knowledge networking research for open science and open scholarship?

What would be the nature of the efficient subsidies for public funding of knowledge networking and its development?

On the basis of what we know about why intellectual or research collaborations succeed or fail, what conditions should be set for how knowledge networking collaborations are organized and governed?

Does a democratic organization necessarily imply that standards in knowledge networking environments are set only at "the lowest common denominator"? What evidence exists for displacement (or replacement) of expertise, "de-skilling", or" re-skilling" in knowledge networks?

Recommendations:

Research on the nature of the relationship between knowledge (having access to it, contributing to its creation, etc.) in networked communications and acceptance of moral, ethical, and organizational responsibility by individuals or groups.

Examples of research suggested:

What is the nature of the relationship between knowledge (e.g., having access to it, contributing to its creation, etc.) and moral and ethical responsibility (e.g., personal, collective, organizational, national)?

If networked environments change as envisioned (i.e., more connectivity, more fluidity, more participants), then may individual responsibility change? Will it be diluted? dissipated? consolidated?

What mechanisms could assist development of individual and collective responsibility, respect for property, or collegiality in a knowledge networking environment?

How can criteria for knowledge networking organization or participation, etc., be set such that they promote (rather than dissipate) responsibility? Examples included certification processes, or requiring acceptance of responsibility before access is granted.

How may these criteria change as knowledge systems become more fluid?

Recommendations:

Research on the establishment of or changes in roles within knowledge networks environments and on how self or group identity may differ in knowledge networking environments.

Examples of research suggested:

How will the emergence of new roles (or the restructuring of old ones)

challenge existing systems of power and authority, both inside research disciplines (e.g., the changes taking place as scientific journals move from print-based to electronic publication) and in all society (with consequent effect on research)?

How can the norms and standards for individual and group conduct be swiftly and democratically articulated in order to avoid error, fraud, malicious and destructive acts, disaster (or will such articulation make any difference)?

What do people value in (from) existing knowledge networks? And how might that experience be applied to larger knowledge networks?

How are identities managed in a knowledge networking environment?

Are boundaries an advantage or a disadvantage in maintaining identity?

How might representation (perception) of a knowledge network within a community influence the community's own social structure? or sense of self-identity?

Compare how the roles of creators, producers, consumers, and users of knowledge are shuffled, interleaved, adopted, adapted ,or denied in knowledge networks.

How might changes in roles affect productivity?

How might changes in roles affect credit and reward systems within academic institution and professional associations, etc.?

How to apply social science theory and data relating to behavior and roles to new knowledge networking situations in which there is an acceptance (promotion) of synergism of roles.

Recommendations:

Research on the development, maintenance, or erosion of trust (among individuals, groups, and institutions) in knowledge networks and how it may be influenced by the nature of networked communication.

Examples of research suggested:

What do we know (or need to know) about n particular, research issues should relate both to the nature of such testbed systems and to the classes of issues that they may be employed in resolving. the establishment, maintenance, and erosion of trust in knowledge networking environments?

How is legitimacy assigned or determined in a knowledge network?

How is trust established in communications without physical presence?

To what extent is physical presence (or the possibility of it) an essential element in establishing and maintaining trust? And a related issue, what does physical presence provide in a knowledge networking?

What mechanism best establishes trust in collaborative environments?

How is scientific authority validated in a networked environment?

What is the relationship of social trust to hierarchies of authority?

To what extent do factors like shared purpose or shared experience influence the development and maintenance of trust?

How may open communication in science be affected by knowledge networking environments, and to what extent is it essential for climate of trust among research collaborators?

Under what conditions are people more (or less) apt to withhold (or contribute) information to a knowledge network? To share their insights with it?

Of what importance is physical presence in establishing communication that participants trust?

THURSDAY

Group 1 -- Issues relating to individual behavior, useability, creativity, innovation

Group 2 -- Issues relating to social/group behavior and to social/political institutions

Group 3 -- Issues related to teaching and learning

Group 4 -- Issues related to ethics, policy, and organization

Group 5 -- Issues related to scientific activities in collaborative environments

FRIDAY

Group 1 -- The Next Generation of Knowledge Networking Researchers

Group 2 -- Relationship of Knowledge Networking to NSF Mandate to Integrate Research and Education

Group 3 -- The Reality of Collaboration -- and How to Foster It at Larger Scales

Group 4 -- Leisure and economics

Group 5 -- Testbeds (group then split into two subgroups):

Group 5(a) -- Resources needed for a testbed

Group 5(b) -- Social and Technological Foundations of Knowledge Nets (for Design)