The system framework and adequacy criteria of automated testing software testing are comprehensive and multi-faceted testing. It is necessary to set up a good system framework, and then after sufficient testing, the software testing work can be completed.
The Framework of Automated testing.
Automated testing has its scope of application and implementation mechanism, it mainly includes the following key elements: the software to be tested, the components and characteristics of the software, the running environment of the software, the external interface of the software, the data input format of the software, the data output format of the software, the capture of the results, the method of automatic testing, the software of automatic testing and so on. According to these key elements, the framework of the software's automated testing system can be determined.
Adequacy criteria for testing automation.
The adequacy criterion of automated test data is to assume that SP is an assembly, SF is a specification set, D is the input field of program P, and ST is a set of test sets, that is, a set of subsets of D. The test data adequacy criterion C is defined as a function: P × F × 7'-{true, false} C (P < F, D) = true, which indicates that the test program P is sufficient according to the specification F and following the test criterion c.
The adequacy criterion describes the testing requirements of the software under test and specifies how to select the test data to meet the software testing requirements. The adequacy criterion is used before automated testing, and the criterion is based on the quality, cost, and schedule of the parties concerned. It is related to software requirements and software implementation but is not immutable. It is embodied in the requirements for the software under test.
In the testing process, not all testing processes are suitable for automated testing. It is necessary to carefully study and analyze the technical requirements and testing steps involved, determine the contents of automated testing in the testing process, and then select mature automated testing tools or develop appropriate automated testing tools.
Although testing automation has become the mainstream testing method, it can not completely replace manual testing. Under some test conditions, manual testing is more efficient than automated testing.
Therefore, before applying automated testing, we should evaluate the stability of the content to be tested. If the test content is not stable and will change at some time in the future, it is not necessary to plan and develop automated testing. Otherwise, the developed software or scripts will not work.
For example, for those time-dependent test content, according to the author's experience, it is recommended not to use automated testing, but to use manual testing. Because the content of the test has a complex time correlation, the automated testing tool developed in a certain period of time is only suitable for the test content of that time period, but not suitable for the test content of other periods. even the test in a certain period of time is only a single test. if you want to carry out automated testing, you must design different automated test modules according to complex problems at different times. Therefore, the workload of developing automatic testing software with corresponding complex time problems is the geometric product of the workload of developing software and scripts without complex time problems, resulting in a large waste of development resources, human resources, and time resources.
On the other hand, manual testing can enable testers to write and modify test scripts flexibly according to the complex time correlation of the test content and test the test content of each time period conveniently. So the author thinks that it is better to leave the tests that are difficult to automate to manual tests.