Test left and right
Familiar testing (also traditional waterfall), is involved in the project received a needs assessment, and then writing a script based on use cases and prepare requirements documents, etc. After developing test lift started testing mention bug, regression testing by the is over, the project to operation and maintenance on the line, then put the next item continues to repeat this process.
This process appears to be no problem, but the drawback is:
- Testing process takes place within a certain time interval, testers have to wait for the product built entirely in order to find errors and faults. Time Admittedly, it takes more than the agreed time can wait for the code to become the bottleneck of testers.
- Testing students are very passive: when the demand for quality, the development of poor quality, you can only passively accept, the result is that you will be a long and painful process because of the poor quality of the test results in on-line extension;
- Bug in the latter part of the cost is very high, it takes a lot of energy and time to repair. Even worse if the product can not be released on time, resulting in great loss.
- After the problem is likely a streaking online for a few months, and finally the business side first investigation was found to lead to a Bug, because of the huge impact of the loss caused to the company for too long, the line was also questioned on why so obviously not a simple question found.
Test test left and right, allowing the test to have more initiative, have more time to test, but not like before because of the high risk of poor quality each time delay on the line, and the line quality of the products can guaranteed.
Whether the test left or right test, they were related to product quality. Do not mention the measure as the start of testing activities, testing activities on the line is the end, not to believe that the quality just test students need to focus on.
First, test the left
Test left thinking, the sooner the nature of the discovery, the lower the chance of unreasonable go wrong.
Test principle left to support the test team early in the software development cycle and all stakeholders . So they can clearly understand the requirements and test cases designed to help software "fast failure", prompting the team to modify all of the earlier bug.
Participation and understanding testers will get the complete product knowledge, thoroughly think clearly a variety of scenarios, real-time design scenarios based on the behavior of the software, which will help the team identified some shortcomings before encoding is complete.
I saw for the test left a good understanding
The real test the left contains two aspects:
1. Improve the quality of ceiling
2. improve the quality of the lower limit
improve the quality of the upper limit, in fact, through a series of activities to avoid problems or let itself become a very good start, a word : a good start is half the battle
to improve the quality of the lower limit, in fact, through a series of activities, so that our quality of the results is guaranteed
Test left, in fact, through a series of activities to improve the quality of the upper limit, shorten test cycles and improve the quality of the lower limit, this way, we can continue to improve in the course of the lower limit, and always will be the quality and stability in a horizontal line while the team together and the pursuit of higher goals.
Never pay attention to code quality of the first day, it planted the bug fixes, the cost and the cost to fix the problem of introducing the issue of re-positioning.
When testing early in the cycle begins, the team will be more focused on quality, and "Let's get the correct coding for the first time" prospects. This helps save a lot of time and reduce the number of iterative software development team must be executed for a particular code.
Test left, what can we do
For the test we actually have a lot left to do, if it comes to the beginning of all for the quality of products and services , then any role in the development process, people have the quality of service.
What activities can improve the quality limit (for example)?
- Health project process (rational and strict adherence to the project processes)
- Reasonable needs analysis (quality assessment of needs, analysis of the reasonableness of demand and integrity)
- Excellent system architecture
- Complete system design (design of quality assessment, analysis of demand rationality and integrity)
- Make full use of static code scanning
- R & D is defined standard
- Earlier test analysis (analysis prior to the completion of the development needs and do various reviews of preparation)
- Early test execution (test execution involved early, we found some problems before integration)
What activities can improve the quality of the lower limit (for example)?
- Health testing process
- Excellent test case
- Reasonable testing program
- The right automation
- Appropriate exploratory test
- Develop self-test (TDD, BDD, test offers better use cases, technical support)
- Quality awareness training team
For the test left, also requires a significant foundation engineering habits, SDLC maturity, hierarchical test, continuous integration, extended release of rhythms on the link, you need to stick to explore the depth of expertise in the field of close of business on the depth, scan the code (specifications, issues, security, exceptions, etc.), CR, submitted by the behavior of the code analysis, test double (mock, fake, stub, dummy), UT, automated acceptance testing. Left needs to have the code engineering efficiency as much as the ability to research and development.
So for testing left, I think we can about the quality of service expansion ideas, participants are not limited to testers
Left practice test questions
Of course, in practice it will be some problems, such as the author team practice together, there have been
- Testing requirements to provide preliminary design, interface documentation! ! !
- Unit testing must pass the test requirements! ! !
- Test design needs intervention! ! !
A lot of people thought that it was required to complete the test in a number of unnecessary things, testing the intervention of my work.
In fact, the problem is that the contradictions I said before the sentence: Whether the test left or right test, were related to product quality. Do not mention the measure as the start of testing activities, testing activities on the line is the end, not to believe that the quality just test students need to focus on.
For the implementation of the test left, the most important is the quality of the full service awareness training
Second, the test right
To the left is the development phase of testing before the shift after shift to the right is released. That it is, after the product line can also do some testing activities. Of course, in a production environment to make a direct test is not recommended, but we can do in a production environment monitoring, online monitoring performance and availability, any problems occur once online, the reaction as soon as possible, early response, giving users a good experience. Technical students than the business side first discovered the problem, the business side has been found that if the volume of business decreased significantly, indicating that the problem is very serious.
In fact, the right test can also be understood as if any problems occur online, the first time we have the ability to identify problems and solve problems, and to ensure the consistency of the data line or as little of the impact of online users, as well as real-time access and customer feedback.
Test right, what can we do
For the test right, online monitoring can be a breaking point, for example:
- Closed loop feedback questions online - check - solve - Update Process
- More convenient log viewer, backhaul service
- Rich valid log, easy to quickly locate problem
- Rich monitoring indicators (such as business outlier indicators)
- Cost control (e.g., SMS, etc.)
- Daily monitoring of key indicators (server metrics)
- Production data monitoring (alarm) (achieved by monitoring production data sql statement, such as whether there are multiple orders the same number of orders occurs, etc.)
So for test right, I think that may surround feedback problems found, positioning, monitoring unfold, participants are not limited to operation and maintenance personnel
Test right, practical issues
Of course the same, in practice it is a problem, in addition to technical problems, there are for example:
- After the unoccupied line monitoring structures
- Online issue feedback mechanisms, business people do not cooperate and so on
- Monitoring indicators unreasonable, but it is believed to increase server load
The implementation of the right to test, in addition to quality service culture, but may be more important: sound feedback, find, locate , after a perfect control framework, how better integration with the project process, do not let it become a burden
Reference 1 https://www.kingwins.com.cn/content-11952.html
Reference 2 https://www.jianshu.com/p/114e041483cc
Reference 3 https://blog.csdn.net/weixin_40180628/article/ details / 79203611
reference 4 https://www.kingwins.com.cn/content-11952.html
