Over the last ten years, I've studied under many teachers and taken many courses. Dr. Brinch Hansen's course was unlike no other. He was interested in solving problems in parallel. I had no idea, even after five earlier courses, that it was so difficult. He took seven to eight different problems and showed by means of a series of beautiful and elegant programs, how one would go about writing parallel programs. His insights were often remarkable, for example, his deep idea that process structures were the correct way to reason and work with parallel processing, just as data structures are the key to sequential processing. Or the time he told us about the importance of constraints in the design process.

Perhaps the conviction always evident in his presentation came from the fact that these programs were his own, and not copied out some standard book. Even now it mystifies me to some extent how he could reduce a really complex program to a series of subprograms each no more than a dozen lines, the whole piece elegantly connected.