Homeless Computer Science Professor

02 October 2005

Learning how to give talks

The current two weeks, 80% of my working hours consist of participating in courses on how to teach. The courses include three days on "teaching in higher education", which covers several aspects of teaching, including giving lectures---and a "Master class" on "presenting with theatre skills". Both are a nice surprise.

I did not expect the courses to be bad, but I was prepared for the worst. What I learnt from the presentation courses I had to take as a student, was that "slides" were holy. Lectures with teachers writing on the blackboard I learnt to associate with complex derivations I never bothered to follow and could better check out later at home. As students we learnt to give smooth overhead or beamer presentations, to which we talked.

I will refrain from discussing now whether the above approach was a good way to get started. Maybe it was. However, in the past years I got convinced more and more that it is not a good way to continue.

An eye-opener for me were Erik Demaine's lectures in the Aarhus Summer School on Massive Data Sets. I can still picture him drawing on the blackboard how to lay-out a tree in memory for optimal cache performance. The nice thing about a good "blackboard talk" is that it is much easier to follow what is happening. You are with the speaker when he writes or draws things on the board. Slides and computer presentations, on the other hand, often contain too much information and the speaker hardly gives you any time to discover the nice bits of information and explanation hidden in the slides.

Another eye-opener was the seminar on approximation algorithms organized by Alexander Wolff for a dozen students in Karlsruhe. The students showed a variety of approaches to giving a talk. Talks given with overhead transparencies or a computer presentation tended to be average; talks with the blackboard tended to be either very good or very bad (the bad student used the blackboard because he "did not have time to prepare slides", read: "did not take time to master the material"). I prepared my own presentation late, during and between the first day's talks, and could thus draw inspiration from the students' blackboard talks while making my computer presentation. The audience was impressed by the way I organized information on the screen.

I was afraid that the course on teaching in higher education would try to convert me back to the powerpoint doctrine: someone must have made that up, it must be the educationalists. Not so. No powerpoint doctrine. It was not them, and the teachers are very well aware of the strengths and weaknesses of the various media one can use. Nevertheless, what they can still teach me about how to give a talk is mostly polishing.

But the Master class provides some insights that are really new to me. To start with, we got explained two models of giving a talk: the clew model and the onion model.

The clew is a logical, linear argument building up to a conclusion at the end of the talk, like a clew unwinding until you finally come to the core. Miss one step in the talk and you lose the plot and miss the point. And yes indeed, this is exactly what happens when I listen to most conference talks and some lectures. In a three days' conference, I actually follow the first five minutes of one or two dozen talks. That is more or less it. After those five minutes, I get lost. I miss a "slide" because I am still thinking about the previous one, get distracted by some random personal associations I had with something the speaker said, or simply doze off because the speaker has a tiring accent.

The onion talk starts with the main message, and adds depth in successive layers around it, always returning to the main message between layers. Since the main message and the main ideas are repeated often, a listener can still follow most of the talk even after dozing off for a minute. Also the talk does not get screwed up near the end when the speaker is running out of time, because by then, the most important things have been said already and the speaker has no reason to hurry.

This is an interesting way to look at it, not only for lectures but also for conference talks. Until recently, I was already prepared to accept that people do not have to remember anything from a conference talk except that they have to remember the speaker or the topic well enough that they may read the paper. This can be achieved in five minutes, so what do we use the other fifteen minutes of a typical conference talk for? We typically use it to explain too much in too little time---all doomed to be forgotten.

I will keep that onion in mind while preparing my next conference talk. I will give one or two talks at the upcoming ISAAC in China, so there you may hear to what extent I decided to go onion and whether it works or not.

So what about the theatre skills, you may ask? Well, this is all so new to me that I will refrain from posting commentaries here. I can confirm though that it is very interesting---and confronting---to have someone look at your expression and relation with the audience instead of just seeing to it that you present a knot-free clew with legible slides. It is too early to tell if I, that is my audience, will actually benefit from it.


At 18 October, 2005 16:32, Jonathan Shewchuk said...

Lovely post. I've found that I get better teaching ratings when I use the whiteboard rather than slides—and more importantly, I can feel that the lecture is going better. The great thing about the board is that it paces your presentation to match the rate at which the students can absorb information, whereas slides usually go by too quickly. Also, it orders the information, whereas with slides students don't know what to read first. I've decided that ordering information is very important: nowadays I design my conference slides so that each slide is broken into several parts, which appear on the screen one at a time as I discuss them.

One evil of Powerpoint is that it encourages slides made of bulleted bits of text. I tell students to always try to express everything in pictures first, resorting to text only where geometric imagination fails them. I use a drawing package (xfig) that makes bulleted text tedious to produce.

If you care to hear another person's opinion, I've written up a short page of advice on giving talks for my students (follow the link on this post). I like your discussion of the "onion," and I think I might quote it on my page.

I would really like to hear about the theatrical aspects of lecturing you declined to write about, or at least see a pointer to some ideas about it.

At 18 October, 2005 20:51, GASARCH said...

(Being the SECOND person to leave a blog
comment it may actually get read!-- I just
left the 17th comment at Scott Aaronsons
blog on a similar topic.)

1) Powerpoint is bad but may be the only
way to teach a large class. My solution would be to NOT have large classes.

2) Powerpoint can be done WELL or

3) Powerpoint caters to the MTV
generation of people with short attention


bill gasarch

At 24 October, 2005 19:20, Jonathan Shewchuk said...

1) I've taught classes of up to 350 in a large room with many large blackboards, extra-thick chalk, and a video camera in the back of the room projecting me and the current board onto a giant screen over my head at 4x magnification. You have to write big, of course. If you don't have those resources for large classes, I guess you might be stuck with Powerpoint.

2) It takes about 5x more time to do Powerpoint well than to do it poorly. So few people do.

3) The job of university is to build long attention spans and a taste for extended thought, 12,000-word essays, etc. Even more important than educating them is to make them self-educating. We need to tell them that there are heights that cannot be attained under stimulus immersion.

4) Alleluia.


Post a Comment

<< Home