Towards measurement of testability of concurrent object-oriented programs using fault insertion: a preliminary investigation

Abstract

There is a lack of methods and techniques for measuring testability of concurrent object-oriented programs. Current theory and practice for testing sequential programs do not usually apply to concurrent systems. An approach towards measuring the testability of concurrent Java programs is proposed in this paper. The key idea is to take a program and insert faults that are related to the concurrency aspects. The approach is based on two methods: (1) mutation of the program based on keywords, and (2) creation of conflict graphs based on static analysis of the code.