An approximate user-complexity measure for relational query language systems

Abstract

Complexity theory provides a powerful methodology for evaluating the time and space requirements of algorithms. It would be helpful if a similar theory could be developed to analyze the level of difficulty presented to the user by database query languages. The relational database languages are of particular interest. Their power and flexibility have real potential for supporting the decision making needs of managers. In this paper an approximate user complexity measure is formulated and applied to three different experiments. These experiments measure user performance in the following tests: (1) relational algebra versus tuple calculus, (2) relational algebra versus domain calculus, and (3) relational algebra versus functional algebra. The results favor the procedural relational languages, and the approximate complexity measure constructed here parallels user performance.