I think enthusiasm is probably the single most important element of teaching. I don't see how students can get excited about material if their instructor isn't excited about it. I love the material I teach (computer hardware and operating systems), and I really enjoy talking to people about it. I try to add humor to my lectures wherever I can use it to illustrate the concepts I'm presenting. If I can make the students enthusiastic too, then they'll learn more.

Unfortunately, given the complicated nature of computer systems today, the surface details often obscure the underlying concepts, and it's easy for a course to turn into a collection of unrelated case studies. I try to focus on the concepts and show students how those concepts can be applied in different situations. I also try to relate the coursework to other areas, both within computer science and in students' personal lives.

The disadvantage of a lecture format is that if I'm the only one talking, it's very hard to figure out what everyone else understands. There are too few exams to use as a gauge of what people have learned. I can sometimes get a feel for the class by watching students' faces (glazed looks spread across the class when I'm not making sense), but I also use several other techniques to get feedback from the class so I can see if they're "getting it" and adjust the lectures if they're not.

One technique is to have students solve problems orally in class. When going through exercises, I'll work down a row of the class having each student in turn supply the next line of a program or carry out the next step in a simulation. I can see immediately how much they understand. There is no way to go too fast for the class, since they supply all the answers.

Sometimes students spend so much energy taking notes that they don't have any time in class to think about what I'm saying. To reduce this problem, I give students a complete set of lecture notes at the beginning of the semester. Although I was a bit uneasy with this idea at first ("they'll know everything I am going to say in advance!"), it had a wonderful effect: everyone in the class started watching me instead of frantically taking notes. The notes are cryptic enough that they don't serve as a substitute for lecture, but they lessen the note-taking burden so students can concentrate on the material.

I try to improve my courses every time I teach them. I'm convinced that if I stop making them better, then they will degrade. Course evolution is particularly important when the underlying technology is changing as fast as it is in computers today, but it also helps to keep me from getting bored with the class.

I continually update my lecture notes and I also gain new insights into teaching through regular discussions with several of the other faculty in computer science. Each of us tries new ideas, and the others steal the ones that work. We also criticize each other's ideas. Criticism is always painful, but it's the only way I know of to improve.

The last ingredient in course improvement for me is to improve my own understanding of the material. In order to find simple ways to explain complicated material, I have to understand the concepts behind the material and focus on them. The more I teach, the more I understand those key ideas, and this in turn allows me to present the material in a simpler and cleaner fashion. As I teach different courses I find similarities between fields, which can be used to help students understand the material.