As we all know, nowadays, whether it is a large factory or a small and medium-sized factory, automated testing is basically a standard configuration . After all, the huge and complicated systems in activities such as Double 11 and 618, as well as the requirements for multi-terminal releases, multi-versions, and model releases, but only "Write some automation scripts" is hard to beat. This can be seen in the recruitment requirements.

However, the reality is that it is difficult to recruit a mature automation test engineer.

Recently, I interviewed a lot of test engineers from big factories: Huawei, Walmart, Tencent, Byte, etc. I thought it was stable every time. I thought that I should have participated in automated testing in big factories, but actually it was a lot of work . 10-year test engineers are still doing functional testing , or mainly focusing on functional testing.

Why is automation testing talent scarce? I have summed up 3 points:

The field of automated testing is limited to the use of tools and frameworks, lacking an overall understanding;
The understanding of automated test design is not deep, and some methods and routines remain in conceptual understanding and cannot be used flexibly;
The value of testing work is underestimated, long-term development is limited, and developers are forced to involve technical tools together.

Do the most cost-effective automated testing

Think first, what is the "end point or value" of our automated testing?

Is it running automatically? This requirement is too rudimentary;
Are the leaders satisfied? Sometimes the project is aborted halfway because of a change of leader;
Is it 100% automated? A high degree of automation does not necessarily lead to high quality;

It seems that it will be difficult to tell for a while, what is the value of automated testing. Until I saw the picture below, it completely overturned my cognition - the final delivery value of an automated testing project is the benefits it generates, that is, the ratio of return on investment to ROI.

At first glance, it is a bit difficult to understand, but if you think about it carefully, isn't that the case.

For example, in the year-end debriefing report, use ROI to express performance : "Boss, the automated test cases I made were used by n scenarios last year, repeated x times, and found y bugs, saving manual work Quantity z person-month".

Is it intuitive? If you want to become a master, you must see and solve more valuable problems and be responsible for higher results.

Become an automation testing master

This method comes from "Former Oracle Senior Development Manager" Liu Sheng 's column "Automated Testing Master" . Compared with the cognitive route from "code ability→tool ability→architecture ability" that 80% of test engineers are familiar with, this new model , and suddenly penetrated the essence of the work of a master tester - to understand business, technology, and architecture, rather than being limited to tools and frameworks.

Compared with the information on the market that only talks about tools, frameworks, codes, etc. like operating instructions, what attracts me most in the column is that the author has created many "automated testing" methodologies that first appeared in the industry ( detailed below), bringing You jump out of the level of tools and frameworks and re-examine the automation test design.

As discussed by the teacher and a classmate in the column:

Now many testers are learning tools on the application layer, and rarely think about problems from the bottom layer and architecture

As a result, they can only write some automated scripts to run, but they cannot understand the cause of the problem itself, especially performance problems. Many testers only know how to run under pressure, but do not know the relationship between the system, network, and applications.

The column involves measurement data analysis, code logic, and job modeling, and also corresponds to data, algorithms, and modeling in software development. He also created a repo on GitHub to put the overall code and related files mentioned in the column . It is convenient for everyone to run by hand.

Of course, although 80% of the content lies in the improvement of "cognition" , he will also list the mainstream tools and frameworks in the industry, as well as selection strategies and implementation practices, and attach a list of full-stack automated testing tools . But this part only accounts for 20%, after all, these things can be found online.

Who is Liu Sheng?

In fact, I have read Liu Sheng's article before, and I was immediately attracted by his fresh perspective. I can feel that he has a very thorough understanding of the testing position, which is also derived from his unique experience:

Worked in Motorola, Oracle China, served as test manager, senior development manager, test director, etc.
Led the team to design the distributed automation framework Automation Center, which filled the gap in Oracle's video conferencing system automation testing field for more than 20 years;
Lecturer at Qcon Global Software Development Conference;

In the daily work of testers, they can only touch part of the "wheel making", and their vision is limited. And practical experts like Liu Sheng can let you understand the testing position from a higher level, which is the most rare thing.

Why is it recommended?

1. Subversion of cognition

Liu Sheng put forward many new methodologies in the column, and this is the first time in the industry:

"Micro-testing Job Model", in this Job model, there is no concept of TestSuite and TestCase, and there is no dependence on specific tools and frameworks, but it is designed for testing requirements and automated testing ROI requirements; it can help you clarify testing scenarios, Workflows, case atoms that require code implementation, etc.;
"3KU matrix" is used to sort out the test capabilities of each section of UI test, interface test and unit test; 3KU test pyramid, each layer of layered test has its own focus, but can realize mutual cooperation and compensation as a whole;
Case DM analysis table, which can analyze different types of automated test development costs and maintenance costs;
Tool four-dimensional maturity model, used for framework selection analysis and decision-making;
......

These are just the tip of the iceberg, which has already opened my eyes and looked forward to it.

2. Systematization of content

Liu Sheng divided the entire column into 4 stages:

The first module takes you to re-examine the basic concepts of automated testing, and master the quantitative thinking method of automated testing benefits - input-output ratio ROI model .

The second module starts from an example of a restaurant ordering system, upgrades from a single application to a microservice cluster, observes the changes in test requirements, and fully verifies the requirements through layer-by-layer testing.

The third module deduces the model design together. Like the design pattern of development, automated test design should also have its own methodology.

The fourth module will provide some measurement models and process examples for driving improvement, and think together how to make a project always observable, controllable, and have feedback. Ensure that the project is always advancing on the predetermined track, and even if there is a deviation, it can be found and corrected as soon as possible.

.......

Finally: The complete software testing video tutorial below has been organized and uploaded, and friends who need it can get it by themselves [Guaranteed 100% free]

insert image description here

Software Testing Interview Documentation

We must study to find a high-paying job. The following interview questions are the latest interview materials from first-tier Internet companies such as Ali, Tencent, and Byte, and some Byte bosses have given authoritative answers. Finish this set The interview materials believe that everyone can find a satisfactory job.