Testing in contest and synthesis of the unknown component: two faces ofthe same coin

In the paper, a test derivation method is briefly described for testing in context when the system components are Finite State Machines (FSMs). In order to alleviate the problem of infeasible machines, tests are derived directly from the embedded component machine as proposed by Petrenko et al. A component FSM can be tested up to the reduction relation of the largest set of permissible behaviors, i.e., the largest solution of the FSM equation computing all valid replacement of the component under test with respect to the specification FSM. As the largest solution usually is a nondeterministic FSM, the methods presented by Petrenko et al. can be used for deriving test sequences with guaranteed fault coverage. However, tests generated by all of the above methods are given in the form of sequences defined over the input/output alphabets of the embedded machine, i.e., over internal alphabets. These tests are then translated into external tests defined over the external observable input/output alphabets of the overall system.