Senior how to determine the standard of software test completion

Catalog

One 、 be based on “ Testing phase ” Principles ：

Two 、 be based on “ The test case ” Principles ：

3、 ... and 、 be based on “ Defect convergence trend ” Principles ：

Four 、 be based on “ Defect repair rate ” Principles ：

5、 ... and 、 be based on “ The acceptance test ” Principles ：

6、 ... and 、 be based on “ coverage ” Principles :

7、 ... and 、 be based on “ Project plan ” Principles ：

8、 ... and 、 be based on “ Defect measurement ” Principles ：

Nine 、 be based on “ quality cost ” Principles ：

Ten 、 be based on “ Testing industry experience ” Principles ：

---

Before software died , If there is no end point for the test , Then software testing will never stop , It can never end . The end of software testing , We should make it according to the specific situation of our company , Can't generalize ! I think the end point of the test is determined by the following conditions ：

One 、 be based on “ Testing phase ” Principles ：

         Every software test usually goes through unit test 、 Integration testing 、 The phases of system testing , We can do unit testing separately 、 Integration testing and system testing develop detailed test end points . After each test phase meets the end criteria , And then the next phase of testing . For example ： unit testing , We require that the test end point must meet “ Core code 100% after Code Review”、“ The function coverage rate reaches 100%”、“ The code line coverage is no less than 80%”、“ non-existent A、B Class defects ”、“ All defects found are at least 60% All of them are included in the defect tracking system, and the defect repair rate at all levels reaches the standard ” And so on . The end points of integration test and system test have related end standards , Of course, it's the same .

Two 、 be based on “ The test case ” Principles ：

         Test designers design test cases , Members of the project team are invited to participate in the review , Once the test case is approved , In the later test , It can be used as a reference standard for the end of the test . For example, during testing , If the test case pass rate is too low , You can refuse to continue testing , Wait for the developer to fix it before continuing . When the pass rate of functional test cases reaches 100%, Non functional test cases achieve 95% above , Allow normal end of test . But when using this principle as the end point of the test , Grasp the quality of test cases , It's critical .

3、 ... and 、 be based on “ Defect convergence trend ” Principles ：

         In the life cycle of software testing, as time goes by , The defect lines found in the test , First of all, there is a gradual upward trend , And then test to a certain stage , Defects are on the decline again , Until the defects found are almost zero or very difficult to find . We can go through the trend chart of defects, the direction of lines , To determine whether the test can be completed , This is also a criterion .

Four 、 be based on “ Defect repair rate ” Principles ：

         Software defects are classified into several severity levels in the test life cycle , They are ： Serious mistakes 、 Major errors 、 Minor error 、 A general mistake 、 Minor errors and test recommendations 6 Kind of . So when we set the end of the test , The defect repair rate of serious and major errors must reach 100%, No functional errors are allowed ; The defect repair rate of minor and general errors must reach 85% above , A small number of functional defects are allowed , Later versions solve ; For small errors, the best repair rate is 60%~70% above . For the test recommendations , You don't have to change it for the time being .

5、 ... and 、 be based on “ The acceptance test ” Principles ：

         Many companies do project software , If you want to determine the end of the test , It's better to test to a certain stage , After reaching or approaching the standards specified by the testing department , Submit it to the user for acceptance test . If it passes the user's test acceptance , You can immediately terminate the testing department's testing ; If the customer acceptance test , Some defects were found , Then you can modify the defect , Submit it to the customer after passing the verification , The corresponding test can also end .

6、 ... and 、 be based on “ coverage ” Principles :

For the test “ coverage ” Principles , In my opinion, as long as the test case “ coverage ” It covers all the software requirements put forward by customers , Including the hidden demand of the industry 、 Functional requirements and performance requirements, etc , As long as the coverage of test case execution reaches 100%, Basically the test can be over . Such as “ In the unit test, the sentence coverage rate is the lowest, which cannot be less than 80%”、“ Test case execution coverage should reach 100%” and “ Test requirement coverage should reach 100%” Can be used as the final anchor . If you don't feel at ease , You have to see how well the test cases work , Check if any use cases are missed , The common functions can be adjusted “ Sample testing ” and “ Random test ”. For coverage in unit testing 、 Integration testing and system testing , Every stage can't be ignored .

7、 ... and 、 be based on “ Project plan ” Principles ：

         Most of the time , Every project is written from the beginning for development and testing Schedule, Accordingly, each milestone will also be corresponding in the test plan , Make a limit on the test progress and end point , Generally speaking, we have to work with project team members ( Development , management , test , market , Salesman ) Reach a consensus , Set a standard end point after the team agrees . If a part of the project is delayed , The test time will be shortened accordingly . In most cases, all the required test contents and regression tests have been run , Can be used as an end point . A lot of nonstandard software companies , It's all about using the project plan as a test end point , But if you take it as an end point , Testing is risky , Software quality is hard to guarantee .

8、 ... and 、 be based on “ Defect measurement ” Principles ：

         This principle may not be widely used , Less understanding . We can do some research on the defects that have been found , Using common defect analysis techniques and tools , Figure it out with a chart , Convenient access , Measure defects over time . I remember before zhuzx I raised the issue of defect analysis technology in this forum , I'm not going to repeat . We can also put “ Defect density during the test period ” and “ Defect density during operation ” As an end point . Of course , The most appropriate test closure criteria should be “ The number of defects should be controlled within an acceptable range ”.

for instance ： How many errors of what severity are allowed in 10000 lines of code , It's easier to quantify , Better to implement , Become the mainstream of testing defect measurement .

Nine 、 be based on “ quality cost ” Principles ：

         A software often has to start from “ quality / cost / speed of progress ” Stop after the balance of the three . As for which of these three aspects is the most important , It depends on what kind of software it is . For instance, ： The software of space and aviation that matters human life , Quality is more important , Even if it costs more 、 Put off the schedule , The test can be terminated only after the test can guarantee high quality , The release . If it's common software , Because of the interests and the market , Even if there is bug, You have to launch the product first , There's no way . Generally speaking , The main reference is ：“ Balance the cost of finding the defect with the possible loss of the defect ”. In the specific operation , According to the actual situation of the company, we can define what kind of situation is “ Testing costs the most 、 The most reasonable ”, At the same time to ensure the maximum interests of the company . If you find bug Cost ratio , Users find bug The cost is still high , You can also terminate the test .

Ten 、 be based on “ Testing industry experience ” Principles ：

         In many cases , Some experience in the testing industry , It can also provide reference for our testing . For example, the familiarity of testers with the industry business , The working ability of the tester , The efficiency of the test will affect the implementation of the whole test plan . If a test team , Everyone has no experience and data accumulation in the project industry , Get a new project , Naturally, I'm confused , I don't know where to start , Naturally, the test quality will not be very high . Therefore, through the experience of the tester , It will also play a key role in confirming the test execution and end point .