Contract-based testing

A tester relies on some sort of contract between the users and the implementers of a system. The contract defines the obligations of each part that are partially checked by selecting and executing test-cases. In fact, these test-cases are contracts, too, although in a rather operational form. In this article we define the general relationship between these two forms of contract and present a method how various test-selection strategies can be defined formally. More precisely, we demonstrate how test-cases can be calculated from contracts using a refinement calculus.