Review

I joined the LDL group at MCC Austin as a part-time member in 1985, while a graduate student at UT-Austin. I was looking for a thesis topic, of course, and had done some work in both databases and logic programming. So I was ripe for the ideas in this elegant paper. The paper showed how the binding propagation achieved by Prolog could be captured through a simple program transformation for a certain class of programs, and suggested that database implementation techniques (e.g., efficient join methods) could be brought to bear on recursive queries.

Over the years, I've returned time and again to this issue, trying to generalize the approach to larger classes of programs, and to understand its implications in the broader context of guiding search strategies. This paper influenced a number of other researchers as well, and today many commercial database systems utilize variants of the Magic Sets idea. For me, it will always remain special as the paper that taught me research could be fun, and simple ideas could cut deep.