Software defect analysis
Author: alan
1. What is a defect analysis?
Defect analysis is essentially the information items contained in the defects, summarizes, summarizes the statistical methods (or analytical models) after classification. The results of defect analysis can be used to measure the quality of work products in software development, understand the area of defect concentration, clear defect development trends. For the improvement of the software process, the release of software products has a very important reference value.
2. Collection of defect data
When we submit a defect report, it is actually recording some necessary information items. In different software manufacturers, the information items that need to be recorded are large. The following is information that needs to be collected when analyzing some data items that must be collected in the analysis defect activity.
1 Defective Severe Level 2 Defects 3 Defects Discovery Time 4 Defects The version number 5 defect discovery 6 is responsible for modifying a defect engineer
u Defects are closed when closing
1 Defect Close Time 2 Turn off Defect Version 3 Repairing Defects Changes Code Route 4 Generates the root cause of defects (for example: demand, analysis, encoding, soft / hard configuration)
3. Analysis point
3.1. Development trend of defects
The development trend of defects includes the growth trend of new discoveries growth trends and closing defects. For software products, the development trend graph is an important basis for assistance decisions. In general, the necessary conditions for software issuance are the increase in the increase in the number of new defects, and see Figure 3-1 below.
Figure 3-1 Trend chart of defects
During the software issuance process, the development trend graph that will help us understand the distribution of defects in each version. Especially in the regression test phase, the distribution of defects can directly reflect the quality of the project.
3.2 Distribution of defects
There are two types of defect distribution, the first is the distribution of defects according to the distribution of the module, and the other is the distribution of the root cause of the defect. The U module distribution map module distribution map reflects the distribution of defects in each module. It can be used to assess the quality level of each module and development difficulty. At the same time, it is also possible to reflect the distribution of test resources in each module from the side. See below 3-2:
Figure 3-2 Module Distribution Map
u Defect generation caused distribution map
This distribution is the most important chart in defect analysis, as it can directly reflect the quality of each software engineering activity, providing direct reference data for improvements in software processes. In general, the stronger the root causes of defects are divided, the more accurate the results of the analysis. See below: 3-3:
Figure 3-3 Distribution of defects
However, when we find more defects in the demand, we can reduce the number of defects by demand review, demand change control in future projects to achieve the purpose of software quality assurance. Also, if we find that there is more problems in the software design process, you can guarantee the quality of the design by strengthening the review activities in the software design phase.
3.3 Source code modification trend and module distribution map
Source code modifications The number of trend graphs can provide reference for regression test risk analysis and software issuance. The more the number of source code modifications, the larger the risk of the code generated by the code, the greater the risk of avoiding risks, the strength of the regression test needs the corresponding strengthening. The same reason, if the number of source code modifications of a product has a rise, it is not suitable for present. All in all, through software defects are not only to be tracked, but also to analyze according to different situations, decrease or avoid defect production as soon as possible, improve software quality.
-------------------------------------------------- -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
[About the Author]
Nette: Alan Liang
Working hours: 6 years of professional expertise: unit test, regression test automation
Hobbies: bicycle photography tennis
Mail: cyber_lover_cn@163.com
