Complexity Measure Based on ProgramSlicing and Its Validation

摘要

The popular single-factor complexity measure cannot comprehensively reflect program complexity and the existing hybrid complexity measure cannot express the interactive behaviors of programs. To treat these problems, in this paper, we propose a complexity measure based on program slicing(CMBPS). CMPBS not only can evaluate factors which affect program complexity such as the length of the program, control flow, data flow and data types of output variables, but also can give expression of the interactive relation between programs. And we also prove that CMBPS satisfies all of Weyuker properties. Compared with the popular complexity measures, CMBPS is a well-structured complexity measure.