Research Summary


Bob Paige's research aims to improve the production rate and the reliability of high performance implementations of nonnumerical algorithms. The approach utilizes program transformations that capture broad algorithm design principles. These transformations are evaluated prior to an implementation by testing whether they can be used effectively both to explain complex algorithms, and also to help design new algorithms. The implementation methodology makes use of conditional rewriting together with logic based program analysis. The program development methodology is evaluated by productivity experiments.

Selected Papers