Matrix-model-based software testing method

Abstract

The invention discloses a matrix-model-based software testing method. The matrix-model-based software testing method comprises the following steps: a) dividing a software system according to function modules; b) splitting and combining the function modules to form layer-level function test points; c) establishing a test case set by structuring a covering matrix for all layer-level function test points, so that each valid value of any pair of input parameters in each level of function test point is coated with at least one test case, wherein in the step c), for the function test point with less than or equal to three input parameters and valid values, the two-coverage test case set is directly established by Latin square. In the matrix-model-based software testing method, by extracting the function test points and matrixing the function test points, the optimal test case set is obtained through matrix calculation, so that the coverage rate, the testing efficiency and the testing accuracy can be improved all around, a testing need of a large software system can be met and the testing cost can be reduced.

Description

technical field [0001] The invention relates to a software testing method, in particular to a matrix model-based software testing method. Background technique [0002] With the rapid improvement of hardware performance, software functions are also showing a trend of diversification and complexity. The complex combination of software function modules has brought unprecedented challenges to test design. How to design efficient and high-coverage test cases has been How to perform accurate test coverage on a huge system in a relatively short period of time has become the primary problem facing the software testing industry. [0003] At present, the main methods of test design are as follows: [0004] 1. Equivalence class division method: the most commonly used method for test design. This method divides the input domain of the program into several subsets, and then extracts a small number of representative data from each subset as test cases. Each input data in the subset is im...

AI Technical Summary

Problems solved by technology

[0013] 1) Low test coverage: In the face of a large number of test function points, the test design based on black-box testing can only cover some combinations, resulting in low test coverage and it is difficult to test the functions of the entire software verify

[0014] 2) The design of test cases is chaotic and there are many repeated use cases: due to the mixed use of various test design methods, the use case design is chaotic, and the verification of the same function points inevitably introduces a large number of repeated use cases. In the actual test execution process However, these repeated use cases not only fail to mine software defects, but take up a lot of test time, prolong the test cycle, and increase the test cost.

[0015] 3) Low test accuracy: the current tests are all generic tests, regardless of primary or secondary, the execution of test cases has no priority division, and the execution mode in the execution process is serial execution, so the test efficiency is low, and It is impossible to make a comprehensive evaluation of the functional robustness of the entire system in a short period of time. The evaluation of software functions can only be performed on the basis of the execution of all use cases, which increases the test cycle

[0016] 4) It is difficult to carry out targeted testing in agile development: the extraction and testing of software core function points is an indispensable link in agile development, and the defects of existing test design methods make it impossible to design for the core function points of software Produce high-coverage and high-precision test cases

For iterative versions, the test coverage is always difficult to achieve the expected results, making it difficult for decision makers to have confidence in the quality of the software, so they miss the timing of software release

Images