Coding compass -- creating a software factory like a flowing cloud

This article is based on CODING Compass Product director Cheng Shengcong Tencent's cloud CIF Sharing at the engineering effectiveness Summit , It has been sorted and updated . At the end of the article, you can go to the official website of the summit , Watch playback and download PPT.

DevOps From the instrumental stage to the process stage

Software engineering from the last century 60 From the 1980s to the present , There is no doubt that DevOps Era , In recent years, the industry has been in full swing DevOps The transformation also confirms this . At this stage , Enterprises have also continued to invest in transformation for so many years , Start eager to see results . Everyone is thinking about a problem , That's it DevOps Whether it really helps business development and digital transformation , Or just the R & D team ？

Assist customers in the last year DevOps In the process of product landing , We are increasingly aware of ： R & D management really can't just rely on building a tool chain , We also need to apply these tools to the actual business processes of the enterprise . We should really reduce the burden of development , Instead of increasing the burden on business development . Only in this way can we effectively improve R & D efficiency , Better meet the needs of business development .

if ,DevOps Before, it was still in the instrumental stage , Various tools emerge in endlessly , So in the context of the rapid development of digital services ,DevOps Is entering a new stage ： Process stage .

Enterprise use DevOps Tools still have challenges

Let's start with a typical user feedback , Let's take a look at the plight of current users ：

The above customer deeply uses CODING more that a year , They have enough say in whether the product is easy to use . By sorting out the feedback results , It can be seen that there are still deficiencies in the products in the instrumental stage . One side , The customer fully affirmed the original choice CODING DevOps Decision , Every role in the team can work on a one-stop platform , The goal of R & D integration has been well realized ; On the other hand , Although our one-stop platform provides the capability modules needed by the team , However, the collaboration between different modules is not well reflected .

1. For products , Its focus on requirements activities is not well related to what development is actually doing , Thus, the progress and risk can not be fully controlled .
2. For development , It is important to update the task status , But because this thing will not block yourself , Whether it is updated in time depends entirely on the level of consciousness . So many times , Developers who are busy with collaborative programming often forget to do this .
3. meanwhile , As a relative post test , Once tested , The inspection of various matters is more boundless , It takes a lot of time to check and update all kinds of information , Plus, there's not much time left for testing , The situation is particularly embarrassing .
4. Not to mention the operation and maintenance colleagues later , I can only repeatedly tell you to be fully prepared before publishing , There is no discount on various verification checks , Then you can only pray that you don't always be in the sensitive release window , There are all kinds of inexplicable problems .

in general , Although different tools on one platform are used smoothly , But from the whole process, I always feel that something is missing . Switching back and forth between tools still takes a lot of effort , And it can't ensure the correctness of the information . All these , These are the shortcomings of tool products .

Enterprises are increasingly concerned about the overall efficiency of R & D Management

This case is not a case , It is DevOps Transformation is a sign of a new process stage ： Enterprises are increasingly concerned about the overall efficiency of R & D Management , From emphasizing the local optimization of a tool , Change to emphasize the global optimization of collaborative processes .

Tools are not equal to overall efficiency , The classical theory of organizational effectiveness management PPT Pointed out in the ： Of an organization 3 Of the three elements ,People、 People are the foundation ,Tools、 Tools empower people , Make your work more efficient , and Process、 Process is the carrier of keeping people's behavior consistent with their goals . There's no point in doing something perfectly that you shouldn't have done , It will even cause damage to the whole . Consider... From a global perspective , A good process is indispensable .

DevOps Products should be built into a new production relationship to further liberate the productive forces

In the Digital Context , The rapid development of business has brought high complexity of software system , Individuals need to deal with more things , It leads to the decrease of single person efficiency . In order to improve the efficiency of each role in the team , Enterprises pursue DevOps The transformation of , Hope to use emerging technologies and tools to quickly improve team productivity . But with more investment in technology and tools , And the team is expanding , At the same time, it also brings the complexity of overall cooperation . These complex dependencies are transmitted to team members like a pyramid , It has formed a great impact on the original working habits and even understanding . Even a simple delivery , All need to go through many operations and the cooperation of different roles , The whole delivery process is therefore fragile and inefficient ： For example, the upstream and downstream contracts and norms are missing , The transparency of the R & D process is not enough , You need to switch back and forth between different tool platforms, and so on .

How to make different tools , Organically coexist in a complete process ？ How to create an efficient process for the team , So that people can smoothly complete high-quality software development , And publish to the production environment ？ In the process , Team members don't have to deal with unnecessarily complex problems , Get caught up in details , Or a long wait delay . We should liberate the productivity of team members , Let developers focus on work that really generates business value . This is something worth thinking about at present ： Just as productivity determines relations of production , We need more advanced R & D management products to empower the R & D team , To meet the needs of today's digital business development .

CODING Compass：DevOps R & D process management products in the process stage

Through to DevOps Sorting out the problems highlighted in practice , We get the following 2 Understanding of three aspects ：

1. At the organizational level DevOps Transformation requires domain experts

7 Issued by the Institute of information and communications in June 《 China DevOps Status survey report （2021）》 Pointed out in the ： near 30% Because of the lack of DevOps Experts lead to slow landing . And when we serve our customers , Also often need to provide advice , Through expert diagnosis 、 Develop a process , Then according to the actual situation 、 Set goals to improve and specific implementation paths .DevOps What the product needs to do is ： Extract effective R & D management experience in the industry 、 And embedded into the product , Guide the customer team to solidify excellent habits 、 Continue to optimize , So as to realize efficient R & D Management .

2. The biggest pain point of team members in collaboration is “ Know everything ”

Based on the existing tools , The team relies on DevOps A simple understanding of , It can be initially coordinated . however , The collaboration problems faced by users do exist ： For example, it lacks the ability to carry out cross functional activities , Lack of collaboration specifications between activities , It is difficult to identify the risks in the R & D process , Individuals need to understand too much context in their work , There are many cross functional operations that can only be handled manually, and so on . These look trivial , But these problems have accumulated and have not been solved yet , It will cause great “ Mental loss ”, It even causes excellent employees to doubt building an efficient organization .

DevOps Deepening development to the present stage , It represents the industry's new expectations for R & D management products ： From agile to DevOps、 combining LEAN The idea of lean thinking , Towards enhanced visualization and traceability 、 Pursue the direction of standardization and efficiency . Based on the perceived pain points ,CODING Combined with their own practice and industry achievements and experience , Efforts have been made to upgrade the product , To help customers better improve their R & D management ability .

Compass = workflow + standard + automation

CODING Created a new R & D process management product Compass, Include 3 Three main abilities ： Namely （ Synergy formed by connecting various activities ） workflow , also （ Standards to improve the consistency of R & D activities ） standard , as well as （ Trigger post activity ） automation . Represents the CODING DevOps In the original DevOps Based on the tool chain , Integrated into Know-how Part of , So that customers can fully learn from the effective practical experience of the industry , Achieve efficient R & D Management .

Compass How to improve R & D management ability

To put it simply ,Compass The product logic is to define the process 、 Standardizing the process 、 Efficient circulation 、 Identify bottlenecks and guide improvement .

1. First , There are various activities in the R & D process .

For example, the product manager will create requirements to backlog Inside , Team development planning will be incorporated into the iteration , And decompose the task 、 Task claim or assignment , Development will create branches 、 Write code 、 Submit merge, etc , Testing is the design of use cases 、 Perform the test , Then the team proposes 、 After passing the quality control and creating a release order, etc .

We know , Some of the things listed here happen within the same role , Some require different roles to work together , In fact, there is a sequence in their progress .

2. secondly , Identify key collaborative activities and concatenate them into a complete workflow .

After classifying these activities according to different roles , You will find that some activities of the same role are objectively the premise of other activities . For example, after the requirements are created 、 It is possible to be included in the iteration , After the branch exists 、 To have the corresponding code submission and MR, Use case design is finished 、 In order to associate the corresponding requirements based on it . These internal relations lead to their activity circulation, which must be completed spontaneously .

For the remaining key nodes , From the perspective of overall R & D , According to the actual work , Artificially define their dependency order , And connect them in series . For example, the corresponding feature branch can only be created after the task is disassembled , With MR、 And the requirements can only be tested after they are associated with test cases , Then perform the test 、 Give the test report 、 Finally, submit the release sheet . This forms a complete workflow .

3. Again , Ensure the robust flow of activities through norms , And the efficient flow of automation driven activities .

In order to ensure the robustness of activity flow , We can set admission and exit criteria for some of these activities , If it does not meet the specifications, it will give a warning and prevent the continuous circulation . For example, the requirements included in the iteration should give the acceptance criteria 、 As the basis of use case design , The pass rate in the test report must meet a certain value to create a release order, etc . in addition , For some activities that can be standardized, created or triggered , You can set automation rules . When the current conditions are met, the flow will be automatic , There is no need for team members to switch to another tool to update the status 、 Or manually create the task for the next step . In this way, an orderly team collaboration workflow is formed .

4. Last , Map the specific steps of R & D to the value flow stages defined by the business , Provide insight analysis .

Specification and automation can produce accurate activity records , So as to provide real and reliable data for efficiency measurement , Conduct effective insight diagnosis and guide improvement . For example, the difference between lead time and processing time 、 Task completion rate / Accuracy, etc . This is value stream management （Value Stream Management） The basis of .

That's all Compass The concept of product design , We want to be able to drive development behavior in collaboration through processes , Let everyone in the process focus on their own value . At the same time, the precipitated process data can accurately perspective the R & D process , And based on the insight analysis of data to guide the continuous improvement of R & D process .

summary

CODING Compass It's based on CODING The original DevOps R & D process management products of tool chain , Including process choreography 、 Process driven 、 Rule constraints and value transfer . I hope it can help the enterprise to get through the goal expectation of the manager and the specific implementation of the R & D team , Achieve the highest responsiveness with the lowest collaboration cost , To maximize R & D efficiency .

At present Compass In internal test , It is expected to open for public beta by the end of the year , Coming soon ！

Go and watch CIF Summit replay