It seems that not a month goes by but that the importance of supporting and fostering collaboration in digital libraries is mentioned. Mike Simmons, who has built, led, and nurtured multidisciplinary, collaborative research environments, has kindly agreed to write for us on this topic.
William McNeill, the great University of
Chicago historian, gave the keynote address at a meeting of the
DOD-sponsored Highlands Forum, held at the Santa Fe Institute
in August of 1996. His job was to set the stage for a wide-ranging
discussion of the future impact of information technology and
the information revolution, and he did so by discussing the history
of information processing in the development of human societies
in an unusual way: McNeill began our discussion of the future
of the information revolution with a description of the prehistory
of dance, extending back into the dawn of our species, perhaps
one million years before the present. When groups of humans move
large muscles together they establish social and emotional bonds.
The result, in the dim past, was to enlarge the scale of activity
from the few to the many, and the bonds thus established served
our forebears well in a whole variety of cooperative ventures
such as the hunt and raiding. Military drill continues to exploit
this technique. McNeill followed the extension of the level of
cooperation in human society by then describing the implications
of speech, ideographic and alphabetical writing, place-value arithmetic,
and printing, up to the current explosion of electronic communication.
The continuing information technology revolution offers new possibilities
for further extending the ability of groups of people to cooperate.
Thus, information technologies take their place in a deep history of human activities and behaviors: meals, family activities, story telling, games, religious ceremonies, and, of course, work. Our word, collaborate, derives directly from the Latin, collaborare, to labor together. The common modern meaning is "to work jointly with others, especially in an intellectual endeavor." The details of how a group works together vary with the field of intellectual endeavor and with the details of the project, but some features are common to many examples of very effective working groups.
The social bonds resulting from movement together, face-to-face discussion, and other aspects of shared activities, even if those activities are not work-related, enable groups of people to collaborate more effectively. In the realm of work, whether we're speaking of food gathering, war, manufacturing, or intellectual endeavor, shared responsibility and cooperation are, we can presume, as old as the species. Solitary hunting is possible -- and even rewarding -- but to primitive man, cooperative hunting parties were an essential means of making a living. Solitary agriculture is possible, but again, the farmer is immensely more effective in cooperation with others.
Cooperative work extends also into the more abstract worlds of art, policy making, engineering design, corporate planning, and research. The accomplishments of geniuses, such as Newton, Mozart, and Shakespeare, may lead us to undervalue collaboration in the creative realms of science, art, and literature. But outstanding individual accomplishments are not the norm. More commonly, progress is made by groups of individuals who either collaborate actively with each other or who build sequentially on each other's work. By fostering collaboration, we enable not only bursts of extraordinary creative genius but also enhance the productivity of large numbers of individuals on whose labors most progress depends.
A modern physics journal, such as Physical Review Letters, will reveal an extreme example: the existence of elementary particle physics teams so large that the list of home institutions occupies a page. Even where teams are not evident, cooperative work exists in the form of inspiration derived from casual interactions and in the dependence of one researcher on the earlier work of others. Making such work available is historically the main role of the library. We now seek to understand how this traditional library role can be extended by the information revolution -- and not only in the form of digital libraries. How may the boundaries constraining the traditional roles of libraries be blurred as new search engines and new tools for collaboration are developed and integrated into digital libraries?
Collaboration is a long-standing interest of mine, perhaps because I see it as a common thread in my past lives in research and management. Nearly all of my research in theoretical physics was conducted in collaborations that depended on cooperative give and take at a chalk board. For many years, I had the fun of being associated with one of the world's premier institutions for research in theoretical physics, the Aspen Center for Physics (http://andy.bu.edu/aspen). The spontaneous and free exchange of ideas about physics among scientists from all over the world is the hallmark of the Center's program: "Although each participant sets his or her agenda for research, few scientists [at the Center] work in isolation and collaborations flourish."
Los Alamos National Laboratory (http://www.lanl.gov), my scientific home for fifteen years, is a multipurpose national facility, which conducts a vast variety of multidisciplinary research and development programs. Of necessity, it is good at promoting collaborations across distance: it is spread over 43 square miles in northern New Mexico and its scientific collaborations reach to every continent. The Center for Nonlinear Studies was created to strengthen collaboration across the Laboratory and around the world in the field of nonlinear dynamics.
The Santa Fe Institute (SFI) (http://www.santafe.edu) fosters collaboration by having a hundred or more researchers visit annually, with about 30 in residence at any one time. On purpose, it has no permanent faculty. For the first ten years of the Institute's research program, my job was to foster SFI's collaborative, multidisciplinary atmosphere which " seeks to catalyze new collaborative, multidisciplinary projects that break down the barriers between the traditional disciplines " In the realm of policy research, multidisciplinary collaboration is essential for many issues but difficult among distant specialists. The Markle Foundation is sponsoring the first steps toward developing a policy-collaboration tool, the HyperForum (http://www.hf.caltech.edu), a Web tool for facilitated, asynchronous, fact-based "discussion promoting discourse on complex public policy issues." In this pilot project involving developers in three institutions and participants around the world, we've continually found that current collaboration tools are far too limiting for the complex task at hand.
In all these enterprises, not only has the best of the work been collaboratively produced, but the management has, at its most effective, been collaborative. This leads me to dream of future collaboration tools and environments that greatly enhance our ability to work together creatively across the barriers of space and time and that respect the hallmarks of a good collaboration: the free exchange ideas, trust, and open oral, written, and visual communication.
As McNeill reminded us, humans have a need to belong to a vivacious primary community to reach full potential. There is good reason to believe that the need for contact with other humans is built into the genome such that most of us can't live successfully and work effectively as isolated individuals. Effective collaboration tools should harness these needs. Our task is to begin to understand how modern tools from information technology might be developed to make "laboring together" more productive while respecting our evolved ways of communicating with each other as humans. Can we build and sustain in the 21st century lively communities of collaborators who are more, not less, effective at cooperation even though their collaboration is electronic and at a distance?
Here are some of the activities and behaviors we see in many productive collaborative groups. Early in the collaboration, the discussion will be highly informal, and ideas will flow freely. These sessions may work in a brainstorming mode, with any idea allowed, criticism minimized, and participants encouraged to add to the ideas proposed by others. Interruptions are frequent. There may be numerous false starts, and the process may appear disorganized. The free exchange of ideas is mediated and controlled by all the subtle signs we're accustomed to giving each other through body position, eye contact, tone of voice, emphasis, and so on. Later in the collaboration, as the task becomes better understood, the work becomes more structured and less wide ranging, parts of the task may be delegated to specialists, and information from related work is collected, analyzed, and integrated. But the informal exchange of ideas is a key feature that remains to the end.
Contrast this with sharing a document on the Internet, using any currently available collaborative tool, or working by fax and telephone. (Hypertext is wonderful, and it makes it wonderfully easy to get lost in a large and complicated site.) The keyboarding constraint is severely limiting. Drawings and other graphical materials are difficult to share, at best, and altering them on the fly is even tougher. Discussing equations and creative approaches to their solution is difficult at a distance. Interruptions are awkward and often irritating and counter-productive. Reference material, even in printed form, let alone in graphical or pictorial form, may be in the hands of only one or a few of the collaborators and thus difficult to share. Indeed, the need to share large and complex data sets among distant researchers stimulated the origin of the World Wide Web at CERN, the great European particle physics research laboratory where the typical experiment involves dozens to hundreds of collaborators. The Web, e-mail, telephone, and fax can keep widely-separated researchers in contact and allow a primitive version of collaboration. As higher band widths become widely available, video conferencing links will improve this in the future. Nevertheless, when collaborators are scattered around the world, and specialized collections of research materials and equipment are located at distant sites, some of our most important research tools continue to be the suitcase and the airplane.
The 21st century medium of collaboration of which we dream will strengthen -- or even help create -- social bonds in the group, especially trust. If we took seriously McNeill's interpretation of the role of dance, we'd seek to find ways to encourage the collaborative group to move about in their local work spaces while the collaborative discussions are underway. Maybe it isn't an accident that we often take a walk together to discuss a difficult point. Collaborative groups often meet around a chalk board or, in the modern world, a white board or easel. People pace around the room and talk to each other; they gesture, they interrupt, they add to what others have written on the board. In scientific exchanges, they change the proposed equations or the approach to their solutions and point to new features and interpretations of data. In engineering, architectural, and other discussions, they add or subtract features by erasing and sketching. My dream of an effective electronic collaboration medium is one that will allow the free exchange of ideas by oral, visual, and written means, help the collaborators access relevant information wherever it's located, and help keep a record of progress.
I dream of a collaboration tool with the following essential features: high-fidelity voice and video communications, support for many-on-many communication (as opposed to the one-on-many communication of TV or broadcast e-mail), flexibility to accommodate diverse styles of work, freedom from the chair and the keyboard, easy manipulation of graphic images, encouragement of spontaneity, ability to support both synchronous and asynchronous collaboration, capability to obtain and display digital library materials for all collaborators to see. Additionally, the tool should incorporate artificially intelligent computational agents, accessible to each collaborator, that enhance the capabilities of an individual, whether working alone or with others, by finding, screening, filtering, sorting, and organizing information discovered on the network for the use of the workers.
Tools for intellectual collaboration should support brainstorming and the free flow of ideas. With permission, I should be able to take the electronic "chalk" from your hand and change the direction of the discussion by correcting, embellishing, or further developing your idea. I should be able to erase or add to a piece of your diagram, your equation, or your list. When we know each other and are in the same room at the same white board, I can see you, hear you, and see what you see. Based on the results of eons of interaction embedded in our genome and our common acculturation, if we're in the same room I may sense how and when to interrupt you and when not to do so. The collaboration flows in a cooperative way precisely because we have the feedback of full social interaction in which to embed our ideas.
Here's a rough description of such a collaboration tool, as experienced by each participant. Its display is large enough that the user can effectively sit, stand, or even pace in front of it. Thus, it might occupy all or part of a wall of your office. In a "virtual reality" or "augmented reality" implementation, the user would move about in a three-dimensional display. The display should incorporate reasonably faithful voice and video communications by other members of the group and, within reasonable limits, should capture voice and picture regardless of the target's position. The tool should display multiple pages of a document under discussion, allow keyboard entry and script entry, including the ability to draw diagrams and equations. It should extend the library by allowing the display of text and graphics from a variety of documents and visible to all. It should offer a faithful dynamic rendering of each participant and a changing map of the pages of any document under discussion with visual links to supporting reference material. This dream is also, of course, a dream about a truly grand distance-learning tool.
Many products are being developed to help with Web site mapping and navigation. Examples are Hyperbolic Tree ( http://www.inxight.com/products/visual/spidview.shtml), Perspective Tree (http://www.inxight.com/products/visual/treeview. shtml), and Perspective Wall (http://www.inxight.com/products/visual/wallvie w.shtml) from Xerox, which help users visualize complex documents. (I haven't used these products and I'm not endorsing them but they do seem to move us in the right direction. I'm sure there are others that I don't happen to know about.) Three dimensional renderings of information spaces are an even more attractive alternative. An example is the work on a prototype virtual 3-D library front end to a database search engine by Worlds, Inc. (http://www.worlds.net) and IBM. Current bandwidth limitations restrict this and other 3-D rendering systems to virtual "avatar" representations of the participants in virtual worlds, not desirable for collaboration.
In an increasingly complex world, specialization will continue, and effective work will require new combinations of skills drawn from collaborative teams of specialists. A powerful collaboration tool should allow easy entry of new collaborators as needed. This means that it should support good automatic record keeping and summary tools so that a new participant can catch up with the group. Our traditional means of integrating new individuals into a group include face-to-face meetings and informal conversations that allow us to react to body language, tone of voice, and eye contact. Even if they won't be able to share a beer or a meal electronically, networked teams may be able to develop trust and close cooperation by electronic means if the medium supports with sufficient fidelity the means of social bonding that are built into our genome by millennia of evolution.
We have not dealt adequately with the problem of asynchronicity. There is no substitute for being present in a discussion in real time. Nevertheless, effective use of our increasingly networked and delocalized working groups demands that good collaboration tools allow membership in the group by those who must participate at different times. This demands a flexible means of capturing the essence of the conversation and displaying where that conversation led the group. It also demands a means for the out-of-time participants to comment on and alter the thread of the conversation as well as its products. As a former academic, I feel comfortable in "leaving it as an exercise for the reader" to grapple with whether this is possible and how.