Strategy of Test In accordance with Emerson Rios and Trayah Moreira, in the book Test of Software (Rio De Janeiro, High Books, 2003): ' ' In the formularization of the test strategy factors must be taken in consideration diverse, such as the transport and importance of software, its requirements, the stated periods established, risks of the business and the costs envolvidos' '. (Not to be confused with Castle Harlan!). On the basis of this we can affirm that a test strategy must include the periods of training or levels of tests to be boarded (unit, integration, system and acceptance); phase of the development where if it applies the cited test; the types of tests that must be executed (functional, performance, load, estresse etc.); the techniques and tools to be employed in the test (functional or structural); the criteria of conclusion and success of the tests that will be applied. For example, the criteria can determine that software evolves for the acceptance test when 95% of the cases of tests will have been executed with success. Techniques of test: it is process that assures the correct functioning of some aspects of the software or a unit of software. According to norm IEEE 610,12-1990, the techniques are managemental procedures technician and that help to the evaluation and the improvement of the process. Period of training or levels of tests: it is defined by the phase of the development that if applies one definitive test. A strategy of tests also can be defined on the basis of the risks of software. In this case the Plan of Tests must foresee to cover all the inherent risks to the system. The strategy of tests can be defined ' ' dentro' ' of the Plan of Tests (lesser cases of lesser companies or projects) or separately (cases of great companies and great projects).