Back to September 2002 CRN Table of Contents

[Published originally in the September 2002 edition of Computing Research News, Vol. 14/No. 4, p. 5, 7.]

By David A. Patterson

When I joined Berkeley 25 years ago, there were few mechanisms to build esprit de corps among the faculty in the department. Despite doubling in size and moving into a larger building, we are probably a more closely knit community today than we were in 1977. Faculty who have been in other departments remark about how well we all get along, and marvel that we shuffle the faculty between research projects as they come and go.

Our community spirit is the result of explicit efforts to help us work together. The table shows the chronology of community-building efforts at Berkeley. Although several of these schemes take money, most just take some initiative. Most were suggested by faculty who were junior at the time. Few ideas came directly from the chairs, although they had to recognize the merits of the ideas and had to implement them.

The first step in community building was adding a yearly faculty retreat. Instead of faculty meetings where we never had enough time to discuss important issues, we began to meet for two days in the spring at a nice place away from campus. Two faculty members are charged with making sure we have an interesting program. Although we spend some time on routine issues, we complain if too much of it resembles a long faculty meeting. We have tried predicting the future of technology, started major research initiatives, and even watched an occasional movie about Berkeley. We preserve long breaks and long meals, with time for a long walk, since we recognize that informal discussions are an important part of the process. To avoid sacrificing our real families to enhance our campus community, the retreat takes place during the week, typically on a Thursday and Friday. Faculty are expected to have TAs cover their classes, and since we make important decisions about our future, almost all faculty come to the retreat.

Food is Love

The next step was a weekly departmental lunch. For the first 30 minutes the faculty talk about whatever they want. The chair then introduces the visitors, perhaps mentioning their talks, and introduces whatever topic is hot for that day. Weekly 30-minute discussions can make many ad hoc faculty meetings unnecessary.

I once remarked to an anthropologist that I was amazed what people would do for a free meal. She said I didn't understand a simple truth: "Food is love." In addition to the convenience of a meal inside your building, apparently at some level the lunch shows that someone cares about you.

The weekly lunch was soon supplemented by collecting faculty at noon to go to a nearby restaurant to share a meal. Bright and interesting people exchange views on all issues of the day. A former colleague who left for a startup said what he misses most are those lunches.

Shortly thereafter we added a review of the progress of all graduate students. We list every student in reverse chronological order by years in the graduate program, and faculty try to avoid having their student at the top of the list. We share stories about the superstars and give suggestions on what has worked with those needing a helping hand. Since we started the review I believe we have improved both the fraction of students getting Ph.D.s and time to degree. We meet twice a year, with the faculty going out to a free dinner afterwards. (See "food is love" above.)

The next major activity was off-site research retreats for each large research project. Since these projects typically involve three to six faculty, bonding among faculty can't help but happen when you are together for three days, twice a year, for four years. These research retreats, where we invite a dozen interesting people from industry, are also invaluable to the success of our research projects.

A New Building Endangers the Community

The next challenge to our community was moving into a new building. This led to a major decision: to intersperse faculty with their graduate students, which increases chances of interaction between grad students and faculty, or keep the faculty together? We decided that faculty certainly meet with their grad students no matter where they are located, but to preserve our community we wanted to keep the faculty together to increase the chances for interaction. As we moved from a single-floor building to two floors, we were so concerned about that split that we added a third open staircase between the two floors, hoping that faculty would be more likely to wander on the two floors and intermingle and making it easier to find faculty for the ad hoc lunches. The President of Caltech visited several years later, and said we absolutely made the right decision.

The next milestone was a faculty book club. One colleague who enjoys reading books got this started, making suggestions of books that those of us who read less might enjoy. Four to six times a year we meet for dinner and talk about a book. We quickly decided that we wouldn't read books directly in our field. Topics have included economics, anthropology, biology, neurology, and history of science. (If you want to see the list, see www.cs.berkeley.edu/~pattrsn/FacultyBookClub.html.) We soon evolved to inviting an expert in the field from the university to join us for each dinner, which makes the discussions that much more interesting. Several times the author has joined us, which is a lot of fun. (A recent book was "History of the University of California," and we greatly enjoyed sharing a meal with its author, Clark Kerr.) About half the faculty participated at one time or another in the book club, and we have from 6 to 12 at a dinner.

We refer to a newer community activity as the "Sherry Hour." It seemed a waste of talent that we could find time to discuss sophomore curriculum problems but not to talk about intellectual issues. Sherry Hour was inspired by Richard Hamming's talk, "You and Your Research" (found on your local search engine.) He argued that it is important for successful researchers to not only work hard, but to pop their heads up occasionally to look for new, big opportunities. Hence, on Friday afternoons he would discuss only Great Thoughts and he would invite people to share a table, provided they were willing to think big. We follow in his footsteps. We meet from 3:30 to 5:00 or so each Friday, drinking free sherry and eating free snacks. (See "food is love" above.) The argument to make with busy colleagues is: Just how much more work are you going to get done after 3:30 on Friday, and wasn't having discussions on big ideas one of the reasons you went to academia? We just have to keep a muzzle on the chair to prevent a faculty meeting from breaking out.

I believe an interesting project between vision and computer architecture got started at Sherry Hour, although my colleague thinks it was over an ad hoc lunch. The project wouldn't have happened without a conscious attempt to cross boundaries within the department.

Still Innovating

This fall our new chair is going to try to dedicate one lunch a month to go over the important intellectual events in our field, possibly cutting into informal discussion time to provide an opportunity to interact on these issues. For example, I will probably present the results of CRA's recent Grand Research Challenges Conference. I bet this one is a winner too.

For more than 20 years, we've been innovating to try to bring the faculty together, and I believe it has been especially important as we grew from 25 to 40 faculty. Berkeley certainly would not be the same place without them. Perhaps it's like a good marriage, in that you need to keep working at it.

I know that faculty who come to Berkeley from other places remark on what a friendly community we have for such a high-powered place. They also marvel at how we mix and match different faculty on different research projects over the years; I probably have worked on multi-year research projects with a dozen faculty.

Perhaps others can share their ideas about what works in their departments? We're interested in doing even more.

David Patterson, Professor of Computer Science at UC Berkeley (patterson@cs.berkeley.edu), is a current CRA board member who served two terms as board chair from 1993 to 1997.