Method of testing programs in a distributed environment

Claim 1

In a distributed processing system comprising a Control Machine and one or more Test Machines, said Control Machine and said Test Machine connected to each other through a network, a method for testing software comprising the steps of: a. storing one or more Test Cases in a central repository, each of said Test Cases comprised of a sequence of one or more inputs, said sequence having a first input and a second input, and each of said inputs having an associated expected output; b. storing one or more error recovery instructions in said central repository, associating each of said instructions with one or more inputs; c. in said central repository, selecting a Test Case to be executed, and sending said Test Case to a Control Program in a Control Machine; d. sending said first input from said Control Program over a network to a first Test Program in a Test Machine; e. executing said first input in said first Test Program, said execution resulting in a first test result; f. sending said first test result from said first Test Program to said Control Program over said network; g. verifying said first test result in said Control Program by comparing said first test result with said associated expected result of said first input if said first test result is the same as said associated expected result, said verification being deemed successful if said first test result differs from said associated expected result, said verification being deemed unsuccessful; h. if said verification is successful, sending said second input from said Control Program to said first Test Program; and i. if said verification is unsuccessful, selecting said error recovery instruction associated with said first input and sending said error recovery instruction associated with said first input from said Control Program to said first Test Program.