Release judgment criteria・Check the test execution rate when judging the amount of test

11/11/202020/10/2021Release Judgment

Check the test implementation status in the test quality at the release judgment

The previous article explained how to determine if the amount of test determined during the test planning stage is sufficient. Actual test, but the it is carried out in accordance with the test plan, from some kind of reason is that in the course of the test can not be carried out or come out the test item, if there is a need added later in or come out test items were , The final test items will be completed as a test report .

In the release judgment , we also look at this test report , compare the planned amount of tests with the amount of tests actually performed, read as information whether the tests were performed without problems, and test. It will judge the sufficiency of the amount of. Father Gutara has determined that the actual amount of testing is okay in the following two ways, and that the final amount of testing is sufficient.

• Is the number of tests conducted significantly different from the originally planned number?
• Is the number of tests performed significantly reduced compared to similar products ?

Be careful if the test fee is significantly reduced

The first point to note is the increase / decrease from the planned test amount (number of test items) . In short , it’s a confirmation that the amount of tests that the test team initially thought they needed and planned was being carried out. In most cases where quality deterioration occurs in software development, there are insufficient man-hours and process delays .

There are various reasons, such as the original development process being unreasonable or an unexpected technical problem occurring, but unfortunately it is common for software development projects to experience some kind of process delay . And, unfortunately, it is a common situation that the release date does not change even if a process delay occurs .

The result is pressure on the testing process, which is the final step in software development . I can’t start testing at the time I originally planned, and even if I start testing, I find a lot of bugs and the test doesn’t go as expected, but the release time is approaching. Such a situation is likely to occur when the amount of tests (number of tests) is reduced by tens of percent compared to the original plan .

Of course, at the time of release, sufficient testing has not been completed, so it will be judged that it can not be released . If the planned amount of tests is sufficient, but the tests have not been performed significantly, it is determined that the test quality has deteriorated due to the insufficient amount of tests.

Be careful even if the amount of tests has increased significantly

Conversely, what is happening when the amount of tests performed (number of tests) is tens of percent higher than the number of test plans? Well there is, the way of software development additional requirements specification if there is a development amount has increased is the case.

As the amount of development increases, the test design will be reviewed and the number of tests will increase. In this way, it would be good if the test plan was reviewed as a result of the increase in development volume, but the release time has not changed , the development personnel have not changed, and only the development volume and test volume have increased . If so, you need to be careful. The amount of tests may be increasing, but the amount of tests required may be decreasing .　

The development and testing teams work hard to keep the release schedule by fine-tuning the order of development and testing . However, as a result, tests to confirm stability, which are time-consuming to implement, may be postponed and left unloaded . Since the number of function / performance confirmation test items that do not take time to implement is increasing, the overall number of test items has increased from the initial plan, but there are also stability confirmation test items that take time to implement. Careful attention should be paid to ensure that it is being implemented as planned. If you don’t have enough stability testing, you need to be careful about whether you can release it .

One way to determine the amount of testing compared to past projects

In both new development and remodeling development, it is logical and good to judge whether the amount of testing is sufficient based on the idea introduced so far, but in reality, Father Gutara has another one. I also used two different judgment methods a lot.

It’s not that difficult, it’s a way to determine if the amount of testing is comparable to past good quality software . After all, software development has various parameters that affect quality in one-item development , so no matter how you think about the theory, there are exceptions, and there is no good theory that can be applied in all cases. Hmm. Therefore, it is quite effective to compare it with the previous case.

If it is a new development, the adequacy of the amount of testing is judged by considering whether the same amount of testing is performed at that time compared to the previous similar product with good software quality. increase.

In the case of refurbishment development, the adequacy of the amount of testing is judged by considering whether the same amount of testing has been performed compared to the product with good software quality in the previous remodeling development of the product. increase.

Well, with the feeling that there is no theory superior to experience, and if a certain number of cases are accumulated, this method will become a fairly reassuring method.

The quality of the test is the product of the quantity and quality of the test, so next is the quality of the test.

Now that we’ve considered the amount of tests performed before the release decision, the next article will show you how to determine if the test quality is sufficient. The quality of the test is divided into multiple articles because it is necessary to think from various perspectives, but please see the article you are interested in.

Next : Release judgment criteria・The first type of test is abnormal and semi-normal