Information and Computer Science, University of California, Irvine, Computer Science 444, Irvine, California, USA, 92717, ackerman@ics.uci.edu
In the old days, we used to sit all in one room around the Mini. Everyone knew what was going on. If I had a problem, I could just ask. Now we all sit in our separate offices [with workstations] and no one talks.. --Astrophysicist.
Despite this promise, the digital library, if improperly considered, could lead to a problematic future. The scientist in the beginning quote inhabits such an environment. Many of the scientists at this research site have spoken with wistfulness about the past, remembering themselves in a single physical room. They do not want to go back; they do not want to lose the independence of separate workstations. But, they do miss the camaraderie and social cohesion furthered by the interaction in that room.
Such interaction is too easily lost and regained only with great difficulty. Considering only the mechanical aspects of access can lead to ignoring the positive and useful social interactions in information seeking.
Second, social interaction is helpful in providing mechanism for seeking informal information. Informal information includes unofficial information such as technical fixes, organizational work-arounds, and personal correspondence [11]. Because this information is often quite volatile and transitory, it seldom written down, let alone indexed. Other people are often the only source of informal information. Current libraries do not handle this type of information seeking, but digital libraries could [2].
Third, information seeking is often ad-hoc and highly contextual [8, 12]. Information seekers often have highly specific interests and needs. It is often more simple and efficient to go to others for information than to written materials. Allen [3], in his classic study, noted that R&D engineers eschewed the formal literature, going instead to colleagues and the trade press for the information they considered necessary. Again digital libraries could provide this functionality.
Fourth, current libraries do have some important -- and useful -- social functions. For students, the university library can have an important socializing function. Students meet one another and talk in the hallways and canteens, and faculty members bump into colleagues in elevators and stacks. Libraries serve as a place to co-learn (e.g., in study groups). Moreover, community libraries offer a number of social outreach and care programs. Many of these useful social functions exist secondarily, as by-products of the library's information access goals. Nonetheless, they are not only useful functions, they also make life more pleasurable and rich.
There is some evidence that such exchanges and interactions provide a "glue" for communities. Fischer [5], while examining people's sense of community in urban and small-town settings, found that the density and public-nature of social interaction may be a critical feature for people's perceptions. Certainly, there is little sense of community in sitting at one's workstation.
The above arguments -- the need for help in selecting material, the desirability of informal information, the ad-hoc and contextual nature of most information seeking, the personal enjoyability and community benefits from social interaction -- all argue for the inclusion of some form of social interaction within the digital library. Such interaction should include not only librarians (or some human helper), but other users as well.
The Cafe ConstructionKit provides a set of reusable objects that include message transport for asynchronous and synchronous communication, parsing for a variety of semi-structured protocols, private and public channels for narrowcast communication, message filters, and message retrieval by a variety of semi-structured methods. The Cafe ConstructionKit is programmable through the Tcl programming language [9][1] By configuring the objects and providing the suitable Tcl program, any application can include the functionality of bulletin boards, chat systems, and electronic mail filters. Additionally, we are actively working on providing the important construction facilities of MUDs [4], so that users can interactively and collectively construct information access methods and environments.
Because of this emphasis on providing building blocks for social interaction, the Cafe ConstructionKit can provide a range of social functionality to digital library applications. It can also serve as a platform for testing various heuristics for interactive information seeking, where users work together to find, create, maintain, and store new information and knowledge.
The current version of the Cafe ConstructionKit is still in prototype. An earlier version exists and was tested on small-scale problems. The first version showed that it was possible to provide a flexible, distributed construction set for interactive communications. However, informal user studies of the first version argued for a better command language (hence, Tcl), using a standard synchronous protocol (hence, the use of NCSA's Data Transfer Mechanism), user interface support for the interactive communication objects, and careful attention to scalability issues. The second and current version is under construction: The core objects (written in C++ and connected through the Tcl interpreter) exist and work, but more work is needed, primarily on the user interface components.
This paper briefly presented a toolkit that makes it easy to add social functionality to applications such as digital libraries. The Cafe ConstructionKit provides auxiliary mechanisms for seeking and locating information without adding significant programming or computational overhead. Most importantly, using the social interaction capabilities of the Cafe ConstructionKit should make digital libraries and similar applications more pleasant to use -- a worthwhile goal.
I would like to especially thank Eric Mandel for his insights about obtaining information and help in scientific communities.