Black Box Testing & white-box testing

Posted on 2009-07-02 filed under: Development (RSS) 129 views

Any work product (Note that any work product) can use one of the following two methods for testing.

Black Box Testing: the function of a known product design specifications that can be achieved for each test to prove whether it meets the requirements of the function.

White-box testing: products known within the working process, you can pass the test to prove whether or not each of the internal operations with the design specifications for all internal components are inspected.

Black Box Testing means testing the software in the software interface Department. This method is to test as a black box, testing personnel do not consider the logical structure of the internal procedures and internal characteristics of the demand for the procedure only on the basis of specifications, inspection procedures in line with the functional description of its function. Black Box Testing is therefore also known as functional testing or data-driven testing. Black Box Testing is to find the following types of error: 1, whether there is any incorrect or missing features? 2, in the interface, enter whether the right to accept? Can output the correct results? 3, whether there is any data structure error or external information (such as data files) visit to the error? 4, the performance is able to meet the requirements? 5, whether the initialization or termination of the error? White-box testing of software is a software to do the details of the process of detailed examination. This approach is seen as a test to open the box, which allows testers to use the logical structure of the internal procedures and related information, design or select test cases for the procedure to test all logic paths. Through screening procedures at different points, and determine whether or not the actual state of the state in line with expectations. Thus white-box testing, also known as structural testing or logic-driven testing. White-box testing procedure is to check the following modules: 1 module of the program execution path for all independent testing at least once. 2, for all the logic to determine, from "Truth" and a "false" in both cases can be measured at least once. 3, in the cycle of operation of the border and within the limits of the implementation loop. 4, testing the effectiveness of internal data structures, and so on. The above facts show that software testing has a fatal flaw, that is, testing is not complete, not thoroughness. Because only a small amount of any proceedings (as opposed to the huge number of exhaustive in terms of) the limited testing found no error, could not explain the procedure does not error.