Tuesday, October 23, 2007


(Even More from Nicole Immorlica, guest-posting from FOCS. This post is from Oct 22, 7:00PM)

I remember taking writing classes in high school. We had a series of assignments emphasizing different writing techniques and topics. But each assignment started with the same question, the question which set the tone for the entire piece of work. Who is your audience? The importance of this question was the most valuable lesson I learned in those classes.

I've now attended about a quarter of the FOCS talks -- all the ones close to my area and a smattering of those completely outside my area -- and it seems to me that there are two types of audiences in every talk. There are the locals, those that are intimately familiar with the research area of the talk; and there are the tourists, those that want to explore something new.

How do you speak to such an audience? Most speakers seem to split the talk into two parts: accessible introduction/overview/problem statement and area-specific implications/proof techniques. And then they have to pack it all into 20 minutes. The result? Minds wander. What can be done about this? Longer talks to allow for a smoother ramp-up to the technical details (and hence fewer papers overall)? Parallel sessions (something FOCS has toyed with in the past)? Or maybe nothing? As my grandmother used to say, "you can please all of the people some of the time and some of the people all of the time but never all of the people all of the time."


  1. Nicole's grandmother is a highly quoted individual.

  2. Still, it's better to be pleasing people (or teaching them) than fooling them!

  3. The majority of the audience is not an expert in your area. The experts already know of your results, and probably know you personally. Of course, it is still your option to ignore your audience and give the wrong talk. That's what some people do, and it is a lot easier.

  4. There is another factor going on here, which is that younger people often feel the need to "impress" the audience (both those in and outside their area) with how difficult their result is. For better or for worse, this is to some extent essential for grad students hoping to get faculty jobs, and assistant profs looking ahead to their tenure letters.

    Once someone gets tenure, they can decide (depending on their personality) whether to keep up this ruse, or whether to aim to present results as simply as possible.