Supporting Verification of the Comprehensive Nuclear Test Ban Treaty Through a Persistent Conversation Interface Stephanie Guerlain 1 , Donald Brown 1 , John Green 1 , Brian Hashemi 1 , Robert Rauschenberg 1 , Christopher Stacy 1 , Roger Mason 2 , and Jane Bohlin 2 , 1 University of Virginia, 2 Pacific-Sierra Research Corporation Abstract Workflow management systems have the potential to make explicit an organization’s implicit need to initiate, track, and execute artifacts and communications about those artifacts. In this sense, these systems can be described as persistent conversation systems, because they are supporting the communication and coordination of distributed actors through a computer- based medium. Furthermore, such systems have the ability to archive and report on current and past activities and communications. We describe the work that we conducted on building a prototype system that has these characteristics. The system is intended to provide an efficient, effective, auditable means of communication between the various groups that make up the Comprehensive Nuclear Test Ban Treaty, thereby enabling the verification of this very important worldwide legislation. 1. Introduction Much research in computer-supported cooperative work and persistent conversations focuses on supporting virtual meetings, virtual conversations, digital chatrooms, discussion groups, and the like (e.g., see [7], [8], [10] and [12]). However, structured , distributed discussions are sometimes an organizational requirement , for tracking and auditing purposes. Thus, an important research question to ask is whether it possible to design a system that helps structure and organize communications according to the organization’s auditing and tracking needs, but does not inhibit effective communications or activities. What this implies is that the structure of the conversational system should implicitly or explicitly inform the distributed, diverse team members of the both the required activities and information bits that constitute a “good” communication. Further, since the intent of such a structured system is to be able to track progress of activities, there needs to be an effective means to browse and monitor the current and past activities. Why would an organization want to manage and track distributed activities? In any distributed organization or system, where information, tools, samples, or other artifacts are passed around, it is necessary to be able to track the status of those artifacts. For example, customer packages that are shipped throughout the world need to be tracked in case a package is lost or needs to be recalled. As another example, blood and other biological samples need to be tracked in a hospital organization. In order to audit and track activities and to collect statistics on the organization’s effectiveness (number of artifacts delivered on time, for example), it is necessary to be able to track the activities of artifacts, and the people’s activities involved with working with those artifacts. Thus, computer-based systems have the potential to aid in structuring activities and communications about these activities in a persistent, auditable way. In this paper, we describe such a system, one that is designed to support international communication and tracking of radionuclide testing samples in support of the Comprehensive Nuclear Test Ban Treaty. Before describing this system in detail, we bring in some theoretical constructs on conversational and workflow management systems. 2. Related work on Conversational and Workflow Management Systems Tatar, Foster and Bobrow [10] discuss computer-based design for conversation and draw on work that was done on conversational theory by Clark and Wilkes-Gibbs [3] and Clark and Schaefer [2]. They describe how a basic unit of conversation consists of two parts – a presentation phase and an acceptance phase – where the two together constitute a contribution. How people establish that they are talking about the same thing is important. This draws on work on establishing a common ground for communication – knowing what your listener knows, and ensuring that each party in a conversation has the necessary knowledge and context to effectively understand the conversation. Thus, a computer-based system that supports conversational activities should help participants to know what item is being talked about, and in what context (i.e., where in the sequence of activities and the 0-7695-0981-9/01 $10.00 (c) 2001 IEEE 1 Proceedings of the 34th Hawaii International Conference on System Sciences - 2001