Philosophy of a colloquium talk: A colloquium should be designed for a general audience of mathematicians, and as such should be much different from a seminar talk. I am encouraging speakers to completely rethink the nature of their talk relative to a seminar talk. Flashing up a slide with ten complicated definitions at the beginning of the talk does not make it suitable for a general audience! Give us the motivation for a definition in addition to (or instead of!) the formal mathematical definition. (What does the definition really mean? Why would someone make such a definition?)
Goal of a colloquium: The main goal of colloquium talk is not for the speaker to tell us about his or her research. I will say it again: your main goal is not to tell us about your research! Rather, the main goal is to explain to a broad audience of faculty and graduate students what your area of research is about. Certainly, you are entitled to tell us (toward the end of the talk) about recent results in the area, but most of the audience will learn much more from a discussion of what your research area is and why one should be interested in that type of problem. Feel free to spend 15 minutes of the talk on a problem that was solved in the 19th century! Telling us about the latest and greatest result should be secondary.
Level of presentation: As much as possible, technical language should be avoided and explanations should be with words or pictures rather than formulas. It is often better to present a simple special case of the problem than to attempt to explain the full machinery of the general set-up. If you work in the area of partial differential equations, keep in mind that an audience member may be a logician who last solved a differential equation in an undergraduate course 30 years ago. Keep it simple! Hide away some of the technical details; maybe lie a little if it gets the point across. Please try to make the whole talk accessible to a broad audience; otherwise, why would a broad audiencce come?
Preparation and length: Since a colloquium is intended for a broad audience, I am asking for a higher level of preparation than for a typical seminar talk. I personally prefer Beamer talks to chalk-talks, since they can incorporate some nice pictures that illustrate the main ideas. (Grab some off the Internet!) Nevertheless, there is a grave danger when using Beamer of going far too fast. If Beamer is used, the number of slides should be strictly limited to about thirty for a 60-minute presentation. Less is more; speakers often vastly overestimate the amount of technical information their audience can actually absorb. (Personal guarantee from the colloquium chair: If you go through 60 slides in a 60-minute talk, everyone will stop paying attention after the first five minutes.)
Organization: Typically, a good colloquium will spend more than half of the talk on background and motivation, before turning to new results obtained by the speaker. In addition, a nontechnical abstract will help encourage people to come. (If the abstract says, "We improve the results of Epstein on quasithin isomorphisms of residually profinite Hecke algebras," nobody is going to come, even if the talk turns out to be great!) In most cases, all proofs should be omitted. If you work in partial differential equations, getting someone in logic or number theory to understand the statement of the theorem is challenge enough; conveying anything useful about the proof is probably impossible.
Please inform your visitor of the expectations of a colloquium and get his or her explicit agreement to this approach before scheduling a colloquium. Talks of a more technical nature should be restricted to seminars rather than colloquia.
Schedule: Typically, colloquia will be on Wednesdays at 4:00, with tea at 3:30.
Procedure: Please first reserve a date with me for the colloquium. After this is done, you should fill out the combined green sheet and seminar form here. (If you wish to fill out only the green sheet only at this time, you can use this link.) On the seminar form, please include some brief biographical data such as where the PhD was earned, name of advisor, current and previous positions held, awards (if any), and major results established. This information helps us get colloquium money for subsequent years, so please fill it in!
The colloquium time slot is generally reserved for job candidates during January and February.