While I was at Sun in the Collaborative Computing (COCO) group, I worked on a range of project that investigated the use of video to support collaboration. I worked on three projects. First, we studied the use of a traditional desktop video-conferencing system. It was the kind of system you build on your first attempt with video: a straightforward "talking heads" video system that was relatively heavyweight, but people really liked it nonetheless. We wrote several papers studying how people used this system and why they liked it. Why Do Users Like Video? Studies of Multimedia-Supported Collaboration is a journal paper that explains what we built, how we studied it with a distributed group, and what we learned about its use. What video Can and Can't Do For Collaboration is a conference paper that looks in detail at some of the conversations among the 5-member team and explains specific ways that video enhanced and detracted from their interactions.

Although people liked the video-conferencing system when they used it, they didn't seem to use it as much as we would have liked, mainly because it was so slow to start up and required a clunky negotiation to establish the connection. If the other person wasn't there (often), the user got little feedback about why the other person wasn't responding. In our next iteration, we focused on making it much lighterweight to start up a connection, and on using the video more to help people negotiate the interaction. We found that the hardest part of getting in touch with others through video was finding the right time to interact. So we created Montage, which supported lightweight "video glances" that enabled people to visually negotiate whether they were open to an interaction, just as people do when they wander by someone's office. Although people again liked the video for the "talking heads" aspect, it was much more valuable for its support of "pre-interaction negotiation." Supporting Distributed Groups with a Montage of Lightweight Interactions is a conference paper about Montage, and the results of our 3-month study of a 10-person group using it. Informal communication re-examined is a more reflective paper that argues for using video to support opportunistic encounters.

In a separate COCO project, we explored the use of video to support much larger organizations. We developed a system called Forum that enabled people to give live talks and presentations to audiences who were watching from their own computer desktops. Audiences saw video of the speaker along with their slides, and they were able to ask questions through audio or text. Forum also allowed the speaker to ask "poll questions" of the audience, and it let audience members send each other text messages on the side. A Forum for Supporting Interactive Presentations to Distributed Audiences is a description of the system we built, the iterative design methodology we used, and a preliminary discussion of its useage. In general, we found that audiences loved the convenience of attending talks from their desktops, but the lack of two-way video made it harder on speakers to feel a connection to the audience. A Comparison of Face-to-face and Distributed Presentations is a description of a more formal study we did to compare Forum presentations to face-to-face ones. We found that twice as many people attended Forum talks than the same talks given face-to-face and the audiences greatly preferred attending talks over Forum, but again, speakers much preferred speaking to a local audience and generally thought they had given better talks when face-to-face. Speakers who wanted to reach a wide audience or who built in a lot of interactive material tended to find Forum the most worthwhile. The paper provides a more detailed analysis of the relative advantages and disadvantages of each style of presentation.

Finally, Studying Video-Based Collaboration in Context: From Small Workgroups to Large Organizations is a book chapter that summarizes our work on Montage and Forum and distills some of the lessons we learned working with video to support distributed interaction. If you want the one-stop-shopping paper on all of this work, this is it.