CONTEXT: Literature reviews

<< Back to main CONTEXT page

Isaacs, E., Walendowski, A., Whittaker, S., Schiano, D., J. & Kamm, C. (2002). The Character, Functions, and Style of Instant Messaging in the Workplace. CSCW ’02.

This study describes in quite detail how people used a prototype application, the Hubbub, for instant messaging. The Hubbub could be run on personal computers and the Palm PDA. During the study over 21,000 IM conversations were recorded among 437 users. 500 of these conversations were then subject to qualitative analysis for determining the conversational characters of IM. Key findings include: media-switching to other media from IM is not very common, most of IM conversations are work related, light and heavy IM-users tend to differ in messaging habits. The study also demonstrates that the availability cues of IM are used by experienced IM-users, and improve the successful initiation of a conversation when compared with other media with no availability cues, such as telephone. (Review by Rantanen, M., 6.5.2003)

Ackerman, M., S. & Starr B. (1995). Social Activity Indicators: Interface Components for CSCW Systems. UIST 95: 159-168.

This paper outlines some of the problems associated with CSCW, such as the “critical mass” and “threshold effect.” It explains why shared representation (e.g. in an aviation cockpit) and social activity indication (e.g. in computer-mediated communication) are important in collaborative work. The study describes the Espresso application, a synchronous chat application that uses social activity indication to inform the user of a potentially interesting virtual conversations taking place. The core of the application is the analysis engine that calculates the relevance of conversations based on user-selected preferences, level of activity and semantic contents. The study proposes alternative visualisations for activity indication, which exploit the relevancy metric calculated by the analysis engine. (Review by Rantanen, M., 6.5.2003)

Vertegaal, R., Dickie, C., Sohn, C. & Flickner, M. (2002). Designing Attentive Cell Phones Using Wearable EyeContact Sensors. CHI 2002: 646-647.

The authors of this paper have recognized that people have sophisticated ways of judging if they can interrupt a conversation without interfering. But this requires visual contact in order for the people to communicate with each other non-verbally e.g. using gestural and postural cues. In telecommunication, e.g. when using the phone, there is limited, if any, way for people to communicate in order to determine if interruption is acceptable. This paper proposes a solution to this problem in the form of a device capable of determining if its user is engaged in face-to-face conversation. This attentive state of the user is determined using a camera, which recognizes eye contact and a microphone that recognizes speech dialogue. If the user is judged by the system to be in a face-to-face conversation, this information is shown to the person that is about to initiate telecommunication. Also, the prototype application enables users to set different notification levels (e.g. vibrate, knock, alert) for calls received in the different attentive states. (Review by Rantanen, M., 6.5.2003)

Cutrell, E., B., Czerwinski, M., Horvitz, E. (2000). Effects of Instant Messaging Interruptions on Computing Tasks. CHI 2000: 99-100.

Previous research suggests that interruptions are particularly costly when the interruption is similar to the ongoing task or when it is particularly complex. The hypothesis in this study is that if the similarity of the interruption to the task has an effect on the interruption cost, or if some tasks are more robust to interruptions, applications, such as IM, can be used to govern the timing and nature of interruptions to optimise user satisfaction. User testing was performed in order to evaluate if relevancy of interruptions to the task or the type of task during which an interruption occurs has an effect on the user performance. The initial results suggest that interruptions that are relevant to ongoing tasks are less disruptive than those that are relevant. Also, task-switching times differ for different types of ongoing tasks. This suggests that IM applications could improve user satisfaction and performance by taking into account the type of task and the semantics of the message, and timing message delivery accordingly. (Review by Rantanen, M., 6.5.2003)

Horvitz, E. (1999). Principles of Mixed-Initiative User Interfaces. CHI ’99: 159-166.

This paper does a very good job in explaining the principles of proactive computing, or mixed-initiative user interfaces. Mixed-initiative UIs bind together intelligent services that are automatically invocated using heuristics with conventional direct manipulation interfaces. The design principles of mixed-initiative UIs are exemplified by the LookOut system that helps the user in some routine tasks associated with calendaring and e-mail. The LookOut system makes real-time inferences about the probability of alternative user goals relevant for this specific application. It then uses these estimates of the user goals to evaluate the utility of providing an automated service. If the utility is perceived high, the system acts proactively and delivers the service. The paper elaborates many important issues related to proactive application development, such as decision making under uncertainty, expected utility of service, alternative interaction modalities and timing of service invocation. (Review by Rantanen, M., 6.5.2003)

Horvitz, E., Breese, J., Heckerman, D., Hovel, D. & Rommelse, K. (1998). The Lumiere Project: Bayesian User Modeling for Inferring the Goals and Needs of Software Users. Conference on Uncertainty in Artificial Intelligence, July 1998: 256-265.

This paper begins with a short discussion on different types of applications that have exploited user modelling. It then focuses on describing the Microsoft Lumiere/Excel prototype whose derivative, the Microsoft Office Assistant, shipped with the Microsoft Office ’97. The Lumiere prototype is intended to help the user in spreadsheet tasks, thus it needs to infer from user actions when the user is in need of assistance to reach his/her goal. The prototype uses a Bayesian Network for modelling the user, and the paper describes very clearly the different design issues that have to be taken into account when inferring user needs under uncertainty. Also interesting is the “Wizard of Oz” design approach where experts are used in the early stage of application development to guess user needs through the “keyhole” of the user interface and evaluate the feasibility of application. The paper also describes the Lumiere Events Language that was developed during the project to enable proactive application development. (Review by Rantanen, M., 7.5.2003)

Van Dyke, N., Lieberman, H. & Maes, P. (1999). Butterfly: A Conversation-Finding Agent for Internet Relay Chat. IUI 99: 39-41.

The authors have recognized that IRC (Internet Relay Chat) users have the problem of finding the conversations that they are interested in. This is due to the absence of hierarchy or organization mechanism to aid a user finding channels. The authors on the other hand do not address the problem of IRC channel activity fluctuation and the need for real-time activity indication (see Ackerman and Starr above). To aid users finding the relevant IRC channels, this study proposes an agent called the Butterfly that visits IRC channels on behalf of users and builds a model of the channel in the form of a vector, and recommends channels that match user preferences. The Butterfly incorporates a user interface via which users may specify preferences regarding the type of content they are interested in, the Butterfly then matches channels to users, in other words, does recommendations for users. The paper addresses the issue of privacy of chat channels, and the authors have designed the Butterfly so that the users on a channel are aware of the presence of the agent, when it is modelling the channel content. The paper also recognizes the problem with private IRC channels, and as future research proposes separate group agents for private channels that invite new participants. (Review by Rantanen, M., 7.5.2003)