This paper attempts to lay a foundation for the mathematical theory of program testing. In six pages, the author only scratches the surface of an enormous problem. The extent of the work is to define when one test method is more powerful than another. This then extends to “programs,” in terms of a test being reliable for a set of programs and specifications. The author develops different categories of test methods, but never produces a concrete example.

The paper is clear and easy to read. The ideas presented are subtle and interesting. My only criticism is the lack of an example.