Software reliability growth models (SRGM's for short) are used to control the software testing process and to make it more effective. This paper describes a family of SRGM's which depends continuously on a parameter. All of the models of this family allow the prediction of lambda0, the failure rate at the beginning of the test and N0, the overall number of faults at the beginning of the test. The models in this family range from "generally optimistic" to "generally pessimistic". The well known and widely used basic execution time model of J. Musa belongs to this family. The new models are the result of a general theory of software reliability growth models developed by the author (to be published). Models of this family are applied to real data. It is described how this family of models can be used for feedback control of the software test process.