A panel discussion is a good way for experts to spend some time rambling in front of a large audience. The audience gets to know the experts in a way that is only possible in person and the expert gets a little more exposure without having to prepare or be reviewed. But sometimes the panel can't get it together and the real expert is in the audience.

'''Therefore:'''
Let anyone in the audience join the panel. Leave one chair open. Let it be known that the empty chair provides the only means for audience participation. Do not take questions from the floor or from the moderator. Do not expect anyone taking that chair to necessarily ask a question, though that is often how their participation will begin.

Make clear also that when the empty chair is filled that one of the current complement of panelists will have to vacate their chair so that there is always exactly one empty chair. Sometimes panelists are eager to leave but cannot do so until they find a subject that attracts someone from the audience to participate in their place. One is welcome to rejoin the panel should they find renewed interest in the conversation.

I've found this is a good substitute for question & answer dialog, especially when the subject is something about change and it is the audience that is going to have to do the changing. I've used the technique with technical people as a way to explore obstacles to change where the subject of change, patterns or extreme programming, is fresh in everyone's mind due to presentation or other workshop activities. I've also used it to "pool the discussion" generated by several authors in a technical seminar. 

I learned the technique from BruceAnderson but understand that he learned it from somewhere else. StevenFraser reminds me that a FishBowl panel is similar. Bruce also talked about a fish bowl as a related technique, but I can't remember how they are different. The park bench is a vivid metaphor, one of a meandering conversation that outlives the coming and going of its participants.

http://www.news.harvard.edu/gazette/2000/03.09/photos/seward.250.jpg

''(Photo from http://www.news.harvard.edu/gazette/2000/03.09/shahn.html)''

The park bench panel runs itself once it gets started. The moderator's job is only to get it started and to confidently remind everyone that silent social communication will make the empty seat rule easier to follow than it might seem. I have one tip of my own design that helps make a park bench start quickly. I ask people to make some notes for themselves and to consult them should the panel's conversation take an uninteresting turn. I'm pretty forceful about how I do this because I want to make sure the shy expert in the audience doesn't forget that he/she has something worth saying.

Here is what I do. First, I summarize what I think the audience knows that I want them to talk about. Then I ask them to take one minute to sit quietly and ponder these issues, paying particular attention to how they might be impacted by the real world in which they live. Then I am quiet for one minute. I check my watch. One minute is a long time when you're doing nothing on a stage. It is even longer when you've just been asked to think deeply. When it is finally over, I give them a specific writing assignment like, write down three things that you just thought about that would be a positive contribution and two things that would prevent you from making such a contribution. This adds enough structure that they know where to start. I'll point out that they don't have to show this to anyone else, but they will need to look at their notes during our next activity, the park bench panel. -- WardCunningham