After working on a variety of collaborative applications, I've come to believe in the importance of making these systems lightweight, in the sense that they allow you to quickly and easily contact someone without a lot of formal setup, messages are transmitted quickly, they allow people to spontaneously "encounter" others, and they allow their users to use their rich set of social mechanisms to coordinate their activity rather than imposing their own protocols. So much of human interaction (in the workplace and in our personal lives) takes place opportunistically when people happen to run into each other, and yet so much of the existing software for collaborative work tends to support much more formal and codified interactions (video conference meetings, conference calls, decision support systems, workflow applications).

Informal Communication Re-examined: New Functions for video in Supporting Opportunistic Encounters is a book chapter that a wrote with Steve Whittaker, David Frohlich, and Brid O'Connaill, which argues the case for using video to support much more lightweight interactions, more lightweight even than our previous project, Montage.

Taking a step away from video, which in itself tends to be heavyweight to install and set up, the COCO group started down the path of developing a system to support a large distributed community that would make it possible for people to encounter others who were doing similar work, i.e. looking at similar data. The system was called Piazza, but we never got a chance to finish it (because of a reorg, you know the story). In concept (and initial implementation), it allowed people to simply notice others "nearby" in their digital world, and then, if appropriate, contact them through an early version of instant messaging, email, an audio link, or (if they had it) video. Or you could simply find out more about the person without contacting them. Say someone sent an amusing email out to your group and you found yourself reading it at the same time as another group member. You could type a line or two to that person making fun of the note, the person might respond in kind, and you'd go on with your work. You just had a quick, lightweight interaction that helped you stay connected with your group. If you happened to be at home while they were at work, so much the better. Maybe the note was about a policy change, and maybe you ran into someone who might know more about it, so you created an audio link to get more details. There was a bunch more to the Piazza idea, which we describe in the conference paper, Piazza: A desktop environment supporting impromptu and planned interactions.

Several years later, when I'd moved to AT&T, I worked on a different type of system that allowed me to explore many of these concepts in a different form. We built something called Hubbub, a mobile instant messenger that focuses on enabling people to become aware of one another when they may be available for an interaction. It makes use of sounds, which can notify you of events when you're paying attention to other things. Each person chooses a Sound ID, a few notes from a song, and when someone become active on a computer, you hear their song play. Hearing them might make you think of something you wanted to say to them, just as seeing them walk by might. Or you might just note it in the background and feel a little more connected to them. Hubbub supports mobile use by running on a wireless Palm (as well as a PC) and by enabling people to move easily between multiple PCs and still get messages wherever they are. It also includes a concept of "sound instant messages," or sequences of notes that have meaning, such as "Hi," "Bye," "Cool," or "Thanks," which gives Hubbub a sense of fun. We conducted a use study as part of the process of building Hubbub, which enabled us to tune the design and feature set to meet the needs of real users. Hubbub continues to be used within and outside of AT&T. We gave a demo of Hubbub at CHI 2001 and presented a full paper at CHI 2002 called Hubbub: A sound-enhanced mobile instant messenger that supports awareness and opportunistic interactions. You can learn more about Hubbub and download it for your own use from the Hubbub web site, www.hubbubme.com

Through Hubbub, we were able to collect thousands of instant messaging conversations, which we used to write a detailed report about how people use IM in the workplace and what they use it for. Our findings are presented in a paper called, The Character, Functions, and Styles of Instant Messaging in the Workplace. We found two styles of IM use: coordinating and working together. Some people used IM mainly to arrange impromptu phone calls or face-to-face meetings, or to give quick updates on activity — to coordinate. Others used IM to do the full range of activities that co-workers generally do face-to-face, i.e. carry out work, discuss work-related issues, coordinate activity, discuss colleagues and their activities, strategize about how to handle difficult situations, chat about personal topics, and so on. Those two styles also varied in the frequency and pacing of conversations.

Also while at AT&T, I worked on a project called ContactMap along with Steve Whittaker and Bonnie Nardi. ContactMap was a research prototype that allowed you to visualize your contacts as a spatial map on your desktop. Each person was represented by an image and their spatial location indicated their social relationship to other people. From each person you could easily initiate contact via any number of media, and you could see what information (documents, previous emails) you had associated with them. Our idea was to reorganize the desktop in terms of people rather than information, which we thought might be more natural and more effective for people. My role on this project was mainly to design the user interface, though I had some input into the concepts as well. We wrote a paper on ContactMap called Integrating Communication and Information through ContactMap, which shows some of the UI.