Statistical Testing using Automated Randomization - Designing and implementing a system for generation of exploratory test cases for an automated testing environment.

The cost of software system testing constitutes a large portion of the total\nsoftware development cost in most projects. High-quality testing requires a well\nstructured approach, thorough planning and experience in the field. The human\nresource cost is often high if a large development project will often employ more\ntesters than developers. Many consider good testing to be both an art and a\nscience, and in this view it is not possible to replace an experienced tester by\nan automatic testing tool.[11] \nThere are at the same time also many areas in testing that are highly repetitive,\nsuch as the repeated execution of a test that has been designed by a tester.\nBy handling the less advanced testing tasks, good testing tools can save time\nfor the tester and both lower the testing cost of the development project and\nincrease the quality of the developed product. \nOne main goal of testing tool research and development is to increase the\nautomation of the parts of testing where a scientific method can be applied,\nand leave the tester free to focus more on those parts that are still considered\nart. This master's thesis will present the design and implementation of a system\naddition to an existing automatic testing tool that will enable the tool to perform\nmore types of software testing automatically.\n