Improving the Collaborative Use of Information Spaces
by Enhancing Group Awareness
J. Alfredo Sánchez
Laboratory of Interactive and Cooperative Technologies
Center for Research in Information and Automation Technologies
Universidad de las Américas-Puebla, México
alfredo@mail.udlap.mx, http://ict.udlap.mx/people/alfredo
 

Introduction

A digital library can be regarded as a virtual space in which scholars conduct research, collaborate and publish their work. In general, available digital library resources include not only information, but also access to a wide range of services and a very diverse community. More often than not, a digital library will pose what can be called the resource overload and resource distribution problems (generalizations of the well-known problems of information overload and information distribution). Resource overload may occur if users are unable to cope with the complexity of (1) a large, dynamic information network, (2) a wide range of services including visualization tools, natural language interfaces, virtual environments and personal and group agents, and (3) communication with users with different backgrounds, levels of expertise, languages and cultures. Resource distribution will be a problem, on the other hand, if some of those available resources are not really accessible to users who could rely on them to satisfy their research needs. These problems can be addressed by providing users with personalized interfaces designed with the community in mind. Communities of users such as those frequenting a digital library represent an enormous potential to assist both experienced and novice users in coping with the resource overload and distribution problems.

The Floristic Digital Library Initiative

We have been working on the construction of a botanical digital library that will support the activities of a highly distributed community including scientists in the US and other countries, state conservation and biological survey offices, federal agencies such as the US Forest Services and the US Fish and Wildlife Service, and the general public interested in plant biodiversity. We refer to this international effort as the Floristic Digital Library (FDL) Initiative. Projects such as the Flora of North America (FNA) and Flora of China (FOC), sponsored by the National Science Foundation, are participating in the FDL. The FNA project alone involves more than 800 scientists working in areas such as the US, Canada and Greenland [Schnase et al. 1997].

We have developed an architecture for FDL that addresses the needs for communication, collaboration and information management among the botanical community. In this architecture, library services are built on top of a distributed object repository comprising a rich mix of electronic documents such as textual descriptions, bibliographic references, maps, and illustrations, which together provide a comprehensive knowledge base of the current understanding of the plants of the world. Services are made available via interfaces that are provided for each type of user and for a variety of activities the library supports. The current architecture of FDL is an evolution of the system architecture presented by Sánchez et al. [1997], in which agent services play a key role. Among the services and interfaces we have developed are role-based views [Tomlinson et al. 1998], agent-assisted taxonomic treatment construction [Sánchez et al. 1998], agent-guided multi-taxonomic browsing [Flores 1997] and personalized information retrieval [Cabrera 1997].

Group Awareness Facilities in the FDL

Users of the FDL are not expected to work in isolation. In order for users to be able to take better advantage of the information and communication resources made available by the FDL, they should be able to communicate effectively with other library users. Users in the digital library need to know about the presence of other users, their level and areas of expertise, and the possible means to contact them. Experienced users in a floristic digital library may want to provide directions or recommendations to newcomers exploring a given taxonomic group. Inexperienced users, on the other hand, may benefit from expert recommendations and knowledge about authoritative opinions. The digital library should make this group awareness possible and should provide the necessary communication mechanisms, but at the same time it should protect the privacy of all users according to their individual preferences.

We have developed Agora, a software component of the FDL aimed at enhancing group (and resource) awareness in the digital library, improving the utilization of available information resources, and increasing the chances for communication and collaboration among virtual groups. This is accomplished by extending existing library services with recommendation services, alerting mechanisms and various communication interfaces.

Users of botanical repositories often get to the information they need by navigating through a taxonomic tree starting with a group of plants sharing a number of characteristics (a taxon) and moving from there to higher (more general) or lower (more specific) taxa. At each point in the tree, users have access to such data as taxonomic treatments, distribution maps, references to authoritative related publications and plant illustrations. Agora introduces three major components that work in conjunction with a taxonomic browser: recommendation services, awareness agents and communication interfaces.

Recommendation Services

Recommendation systems promise to help users in utilizing otherwise unwieldy information spaces and connecting them with like-minded users [ACM 1997]. On the FDL server side, Agora introduces an extensible collection of recommendation strategies available for use in client interfaces. The two basic forms of recommendation, content-based and collaborative recommendation [Balabanovic and Shoham 1997], are supported. For example, recommendations can be produced so the user can be referred to new library elements which are similar to those requested in the past. Similarly, users can be advised to examine library items other users with overlapping interests have requested. Recommendation services operate in close relationship with a taxonomic browser, since much of the success of recommendations depends on following the users' actions closely as they traverse the FDL. Agora allows users to define an interest profile that can be used to find other users in the library with similar interests. Experts in each taxonomic group can be identified and can provide and rank explicit recommendations for users browsing specific library areas.

Awareness Agents

As a user navigates through the FDL, he or she may access items for which a virtual group of users has already been formed, an expert or the author herself may be logged in at the time, or an explicit recommendation for related items has been made. Awareness agents bring these events to the users' attention and allow them to take immediate action to contact other users or groups of users or to access available recommendations.

Communication Interfaces

Indeed, a set of possible actions for the user in the events described above is enabled by the third component introduced by Agora, group communication interfaces. FDL users are allowed to initiate or participate in synchronous or asynchronous individual or group communication. For example, if a scientist who is an expert in a given genus is present at the library while a student is examining that particular taxon, and the expert has specified in his profile that he is willing to take questions or engage in a live dialogue, then the student is presented with the option to contact the expert either synchronously (e.g. via an audio channel) or asynchronously (e.g. via email). Similarly, if a virtual group exists to discuss the item in question, the student is presented with the option to join the discussion, either synchronously (e.g. via a virtual room) or asynchronously (e.g. via a mailing distribution list).

Prototypical Implementation

Building on existing FDL browsing functionality, we have developed a prototype that demonstrates the three components of Agora described above. Newly developed functionality has been implemented as C programs that access the FDL database via the CGI interface, Java applets and JavaScript programs that display new information and communication options as independent windows containing textual and graphical elements. A snapshot of the browsing interface along with alerting agents and recommendations presented to the user is shown in Figure 1.

In Agora, users rank their recommendations of library elements (from unrelated to must see) and also add annotations so other users (or their agents) can filter this information based on their needs and preferences. Although authors of electronically published materials are automatically considered "experts" in the taxonomic area with which the material is associated, all library users can keep their own profiles, specifying their perceived level of expertise.

 

Figure 1. FDL's taxonomic browser, Agora's main panel, and a recommendation volunteered by a user agent named Wogant.
 

Implications for Collaborative Information Seeking

Much work is still ahead to evaluate and refine the design of Agora. The FDL itself is currently under development, so testing in actual library use scenarios is still pending. However, initial positive feedback from users indicates that this development will contribute significantly to assist users in using more fully the library resources (information, software services, user community), discovering otherwise inaccessible information and relationships among library elements, and dealing with a highly complex and dynamic environment.

Acknowledgments

The current prototype of Agora has been developed by Lulú Fernández. Support from the Center for Botanical Informatics of the Missouri Botanical Garden has been instrumental in the development of this project. Part of the work described in this paper is funded by grants from the Andrew W. Mellon Foundation and the National Science Foundation (DEB-9505383 and DEB-9626806).

References

ACM. 1997. Special issue on recommender systems. Commun. ACM 40, 3 (March).

Balabanovic, M., and Soham, Y. 1997. Fab: Content-based, collaborative recommendation. Commun. ACM 40, 3 (March), 66-72.

Cabrera, J. 1997. Integración de servicios y agentes de usuario en la recuperación de información en una biblioteca digital. Tech. Rep. ICT-97-2. Laboratory of Interactive and Cooperative Technologies, Universidad de las Américas-Puebla, Cholula, Puebla, México. (http://ict.udlap.mx).

Flores, C. A. 1997. Agentes de usuario como guías en bibliotecas digitales. Tech. Rep. ICT-97-3. Laboratory of Interactive and Cooperative Technologies, Universidad de las Américas-Puebla, Cholula, Puebla, México. (http://ict.udlap.mx).

Sánchez, J. A., López, C. A., and Schnase, J. L. 1998. An agent-based approach to the construction of floristic digital libraries. To appear in Proceedings of the 3rd ACM International Conference on Digital Libraries (DL'98, Pittsburgh, PA, July), 210-216.

Sánchez, J. A., Leggett, J. J., and Schnase, J. L. 1997. AGS: Introducing agents as services provided by digital libraries. In Proceedings of the 2nd ACM International Conference on Digital Libraries (DL'97, Philadelphia, PA, July), 75-82.

Schnase, J. L., Kama, D. L., Tomlinson, K. L., Sánchez, J. A., Cunnius, E. L., and Morin, N. R. 1997. The Flora of North America digital library: A case study in biodiversity database publishing. Journal of Networks and Computer Applications 21, 1, 87-103.

Tomlinson, K. L., Spasser, M., Sánchez, J. A., and Schnase, J. L. 1998. Managing cognitive overload in the Flora of North America project. In Proceedings of the Thirty-First Hawaii International Conference on System Sciences (HICSS-31, Kohala Coast, HI). IEEE Computer Society Press, Los Alamitos, Calif., 296-304.