Ka-Ping Yee, 2003-09-22 (revised 2003-12-03)
Design, implement, and evaluate a discussion system for groups of up to 30 motivated participants that addresses the following three aims:
Aim 1. Assist in the organization of knowledge and opinions surrounding a design or policy question.
Aim 2. Facilitate the formation of consensus around proposed solutions.
Aim 3. Support the retention and ongoing reuse of knowledge from discussions.
Face-to-face meetings and online meetings each have their own significant advantages. Face-to-face meetings provide a much richer and more expressive communication channel, capable of conveying emotion and deixis, allow for the shared use of tools and props, and facilitate direct collaboration on work pieces such as documents or design artifacts. However, such meetings can be difficult to schedule, and travelling to meet in person is expensive and time-consuming. In addition, because most people can only focus attention on one thing at a time, meetings don't scale well. Participation must decrease as the number of participants grows; it is difficult to hold a productive face-to-face design discussion with more than a dozen people participating. Online meetings (particularly asynchronous discussions) provide the possibility for many people to participate without having to take turns. Another commonly cited advantage of chat rooms and mailing lists is the existence of a lasting record of meetings and messages. The record enables knowledge transfer of the meeting to people that weren't present, and can provide increased value with the addition of a search facility.
The participants in a meeting generally arrive with different perspectives on the issues. Differences in opinion may stem from a wide variety of sources, including differences in fundamental values, exposure to different sets of evidence, differences in interpretation of the facts, differences in ultimate goals or priorities, and so on. The process of making a group decision is in large part an exercise in sharing these perspectives to generate a common understanding among the participants. Lack of structure in expressing opinions and perspectives can make it difficult to identify common ground and sources of difference. Thus Aim 1 is to support the construction, exchange, and collaborative development of frameworks of knowledge surrounding a decision.
Moderation can take place in both face-to-face and online contexts. However, it is interesting to note that moderation techniques and procedures are much more sophisticated and more often employed in face-to-face meetings. One possible reason for this is that face-to-face meetings have been taking place for millennia, whereas online meetings are a recent phenomenon. But this is not sufficient to explain why there has been so little transfer of face-to-face moderation procedures (such as Robert's Rules of Order or consensus voting procedures) to online meetings. Therefore, Aim 2 of this proposal demands an examination of the factors that make certain moderation procedures more or less commonly used in online meetings, followed by an attempt to transfer procedures that are known to be successful in real life to online discussions, or to design new procedures that work for online discussions.
A persistent problem of face-to-face meetings is the inability to retain lasting value from a meeting after it has taken place. Diligently taking notes can distract one's attention away from participating in the meeting, and even meticulous notetakers can still miss important information because the knowledge of which information is important may be available only at the time of retrieval, not the time of recording. No doubt many of us have had the experience of going to a conference and taking copious notes, only to never look at those notes again. The problem is that after the notes are recorded, they retain less value out of context, and are often recorded in a form that is difficult to reuse. Lack of reuse dooms us to repeat the same discussions or even the same mistakes again and again; ideally, we would like to reuse the effort expended and the knowledge gathered from past decisions to help guide future decisions. This requires both support for retaining information in a useful form and for locating and reusing relevant information after it has been stored. Aim 3 states the necessity for the retention and reuse of the supporting evidence, rationale, and other information behind a decision.
In addition, part of the motivation for this work is to advance the state of the art towards electronic governance and to increase citizen participation in democratic decision-making. Although this is a distant rather than immediate goal, the retention of a public record of a debate is helpful because it promotes the accountability of government representatives and encourages them to represent the interests of their constituents. Future or related work, not directly in the scope of this research, might look at providing greater involvement for citizens by letting them give feedback to their representatives during a debate.
It is an assumption of this work that the participants in a discussion are not hostile or malicious toward the process itself. Although the participants may have different motivations and opinions, we require that they share an interest in generating an outcome. A group of 30 or fewer participants should be able to handle exceptions to this assumption by electing a moderator to eject members that are actively disrupting the process.
Many groups of people already conduct their business and decision-making online, using real-time chat, e-mail distribution lists, or other groupware. This work will draw on the historical record of successful and unsuccessful attempts to structure decision processes.
Several open source development teams have semiformal procedures for voting and decision-making on mailing lists. The following projects use procedures based on those initially adopted by the Apache HTTP Server Project. These procedures use textual conventions in e-mail to convey votes.
As explained in an article about conducting board meetings online, the Society of American Archivists has experimented with applying mailing lists to governance. Although the author cites advantages such as the speed and ease of distributing relevant documents, reducing telephone expenses, and the increased responsiveness of the organization. Surprisingly, she does not mention the benefits of having a stored record.
The nonverbal cues in face-to-face encounters are missing from e-mail messages, so "listserv deliberations can unnecessarily prolong the decision-making process." Because of the lack of the persistent common context that would be supplied by an agenda, presentation, or whiteboard in a physical meeting, "the one-dimensional nature of online debate can create a bureaucratic nightmare. It's no simple feat to keep track of motions, amended motions, and multiple discussions." An interesting observation is that having too much information can cause problems: "Decisions that seem simple can become complex as unexpected information surfaces in an online discussion." The author finally suggests that "What SAA ultimately needs is a new parliamentary procedure to embody this new form of governance."
The SAA has established a set of simple conventions for keeping track of motions and votes using tags in the Subject line (such as "CB:" for Council Business), and formal procedures for the voting process.
Crit is a service permitting public annotation of public Web pages in any browser. The stated goal of Crit was to give voice to dissenting opinions on published documents such as news articles, in the hope that this would help to weed out errors or misleading statements. Crit was a technical success in that it achieved its functional goals of providing annotation across the broadest possible set of target documents while maintaining maximum compatibility with the installed base of Web browser and server software. However, it was not a practical success due to its lack of a specific workflow design target.
Crit is relevant to this work because it is a step towards a tool for the citation and critique of external information sources, which provide the foundational evidence upon which policy discussions are built.
Zest is an e-mail processor that converts the messages in a mailing list into a browsable Web page with an organized summary of the conversation. It uses an unusual but simple algorithm to thread the messages together and mark lines of conversation, and permits the participants in the discussion to mark up their messages with simple typographic symbols indicating their relationship to other messages.
Zest is relevant to this work because it is an example of how to extract a rough knowledge framework from a discussion in a very low-cost way.