On Program Dicing

Abstract

Since it is sometimes difficult to apply the technique of static program dicing in debugging programs, we introduce dynamic dicing as the dynamic counterpart of static dicing. As the effectiveness of program dicing techniques depends on the size of the program dices used, this paper uses a probabilistic approach to investigate the relationship between the size of static and dynamic program dices. In making a comparison between dynamic and static program dicing, we present and prove six propositions. This leads us to the conditions under which a dynamic program dice is smaller than a static program dice. Based on those findings, we offer five strategies for constructing dynamic program dices. © 1997 by John Wiley & Sons, Ltd. J. Software Maintenance 9: 33–46, 1997