A typical software testing is performed by making and running test cases that consist of input pattern and expected output pattern. Because exhaustive testing requires billions of test cases and many labours, we can only perform hundreds of them in practice. As for a software logic testing, selecting test cases from many candidates to achieve enough programme path coverage is important, and specific software testing skills are required for it. In our study, we propose a knowledge creation method of software logic extracted automatically from programme source code. In our method, all possible programme paths are extracted from source code, then converted into a decision table, which is easy-readable table format for software testing engineer. The logic verification can be performed exhaustively in a short time by comparing the decision table with a specification of software. Our method would contribute to improve both efficiency and quality of software testing.

No abstract

No abstract

No abstract

No abstract