An efficient approach for logical ripple effect analysis
Y.-C. R. Chen · 1987
It has become apparent that software maintenance cost dominates the total expense of computer software. Among many activities of the software maintenance, identification of the logical ripple effect and elimination of inconsistencies during the software modification are of significant importance. Logical ripple effect analysis can be applied in two aspects: to identify which part of a program that must be checked or modified to insure the consistency of the initial modification, and to estimate quantitatively the program quality such as logical stability which represents the resistance to the potential logical ripple effect. The existing approaches do not accommodate both identification of the logical ripple effect and estimation of logical stability at the same time, while in this approach, the technique can be used for both purposes. This approach is more accurate in calculating the logical stability as well as more efficient in identifying the program areas which need more maintenance effort. Further, this approach is applied at the source code directly while the existing techniques are based on the graph models of the program. The consequence is the saving of execution time for this approach because many syntax details can be ignored due to the fact that the program has been compiled and is syntactically correct in the maintenance phase. This approach may be also used for program debugging such as error correction in the maintenance phase. Overall, the technique may be employed to evaluate the logical stability of programs, and to trace the program areas which affect where errors are observed. Finally, the technique can also be used to identify the program areas which need further examination after program modification to insure its functional consistency. The approach has been implemented and experimented for identification of logical ripple effect and calculation of logical stability.