For many years, we had struggled to obtain the institutional support and resources to develop advanced courses where hands-on experimentation could be taught. This is a much more difficult task than developing and teaching a course where only lectures are involved: it requires sophisticated experiments that work right when the students try them, usually a task which requires months of dedidated effort for which it was impossible to find the time or support from the administration in our hectic environment. On our second attempt, we won the NSF projects. In 1992, our research team used a Macintosh Quadra computer purchased under the NSF project to incorporate a video image processing assignment in AE3010 (Step 1). Students from this course formed a team which developed more experiments, and then signed up for a course on flow diagnostics, AE4010, being taught for the first time. In the course , 3 teams each took on the development of one experiment, and the manuals to go with it. By the middle of the quarter, the experiments had started working, and the teams had learned much of each other's experiences. Each team then became the "guides" to help the others use their experiment. These experiments and techniques were used next quarter to develop a course on Flow Control. The results from these courses and from research experiments were used to develop image-based assignments and literature data bases.
Today, these courses have become "routine", and image-based problem-solving techniques have been used with success to improve several courses. The key feature of these courses is that we learn to do new things by iteration: that is, we stay with the problems through several attempts until things really start working, instead of being limited to one set of hit-or-miss exams. This is how we learn things on the job, and the new technologies allow us to do these things within the constraints of a high-pressure engineering curriculum.