On convergence toward a database of program transformations

Abstract

Several fairly large sets of programming rules have been developed recently. It is natural to ask whether the process of developing such rule bases may converge. Having developed sets of rules for specific programming tasks and domains, will they be helpful when other tasks and domains are considered? While it is too early to give definitive answers, experience with the rules of the PECOS system has been positive. Both during the process of developing the rule set and while developing rules for another domain, the existence of already codified rules proved very helpful.