Advice on Panel Proposals

• The topic: Broadly speaking, there are two kinds of panel topics: (1) very controversial (e.g., the relational/OO debates of the late 80's, the role of XML in data management) and the (2) investigative/informational topics (e.g., research issues stream processing). The former tend to be most memorable and entertaining, but such topics are increasingly hard to find these days. The latter can be very valuable, but special care should be taken that they don't deteriorate into individual research presentations. Even for such panels, a good topic is one in which there are significant differences of opinions.
• The participants:
  • Limit the number of participants to at most 5.
  • Make sure all sides of the controversy are represented. Often, people outside the specific area of the panel are very valuable and provide unique insight.
  • Balance the panelists between academicians, practitioners (note that working for an industrial lab does not make one a practitioner).
  • Getting practitioners on the panel is often quite a lengthy process (yes, they are busier than academicians!). Start early.
  • A mix of nationalities on the panel is also recommended.
• Running the panel:
  • Prepare a list of questions for the panel to address. Have some more questions for the discussion following the presentations.
  • Get the panelists to tell you their positions in advance, so you can ensure balance and controversy. It's usually a good idea to meet and coordinate during the conference.
  • As moderator, make sure to stop panelists at the end of their alloted time. The most common mistake is to let panelists run over time, and therefore not have time for discussion at the end.
  • Remind the panelists that they're not there to give shortened research presentations.