How would I work as a technical manager… 1
1 blanket… 1
1.1 the introduction… 2
1.2 Tasks of the Technical Manager (Director)… 2
1.3 This position is highly suitable for me… 2
2 Terms, concepts, assumptions… 4
3 planning… 5
3.1 Different core elements in different stages… 5
3.2 Organizational structure of R&D Department… 6
3.3 specification… 8
3.4 Concerns… 9
4 plans… 10
5. Training and Assessment… 11
5.1 training… 11
5.2 the inspection… 11
5.3 My assessment… 12
5.4 the promotion… 12
6 Requirements and Testing… 12
6.1 The pros and cons of different characters in different stages… 12
6.2 About demand personnel… 13
6.3 About Testers… 14
6.4 Confirming or Sorting Requirements… 14
6.5 Requirement Decomposition, Task Tracking… 14
7 Single code walk check… 14
7.1 Development stage of code checker… 15
7.2 Benefits of Code Lookup… 15
7.3 How Long does it Take me to check the code… 15
8. Specialized division of labor and establishment of the company’s public library… 16
8.1 Specialized Division of Labor… 16
8.2 Setting up the Company’s public Library… 16
8.3 Technical Breakthrough… 16
Since its establishment, The company has been in a state of rapid development, and will develop rapidly in the next few years. Rapid development means: easy to fall. Fell down, good luck, continue to run, grab some leftovers cold juice; Bad luck, straight out. They had to run again, and when they had finished, they found no way to run: the incremental market had been robbed.
Introduce the computer game “double point hospital” classic plot: “the profitable hospital suddenly huge loss, the symptom is: each department is overcrowded”. The new man’s solution: build more departments and hire more doctors and nurses. The new doctors and nurses were new and inefficient, making the congestion worse. The root cause: intractable diseases take up too much of staff’s time, and those who are difficult to treat (less than 10% of patients) are sent back, quickly turning a loss into a profit. In software development, it’s less than 10% of the requirements, leaving the team running out of steam and losing money.
In reality, this problem can not be solved by the director of each section and the director of each type of work. If the hospital has a CTO, it’s the CTO. Most hospitals have no technical director, so they can only be solved by the chief and deputy directors. If the directors are technical, managerial, marketing, and not financial, there will be no solution.
Core tasks of technical manager (director) : Progress and quality control of the whole process of product r&d, including:
Adjust the department structure according to the core demands of different products and stages. Allocate resources (mainly manpower) among departments, products and jobs, and request resources from superiors. Coordinate product positioning to serve different levels of users and analyze whether there are faults. In view of the high incidence of problems, the establishment of adequate systems and norms. Handle inter-departmental conflicts and coordinate inter-departmental cooperation. Daily tasks
Communicate with the company, understand the company’s business, strategic direction, and make corresponding plans. Familiar with management knowledge to facilitate communication with superiors. Communicate with other departments for mutual cooperation. Understand market knowledge to improve communication efficiency. Review code, requirements documentation, and test results to keep abreast of the employees involved. Continue to supplement the requirements knowledge, familiar with the main product business knowledge, understand the secondary product business knowledge. Examine the results of various departments (departmental self-inspections are often incomplete) and identify bottlenecks. Communicate with requirements analysts, pre – and post-sales (and even customers and users) to estimate the effect of requirements. Categorize or summarize defects based on their cause to identify development team or even individual bottlenecks. Compare the defects found by code walkthroughs with those found by testers to estimate bottlenecks between the two. Process applications from subordinate departments. This position is highly compatible with me. The key to “control the progress and quality of the whole process of product research and development” is to read and write documents. Can read the document, can make the right decision; Ability to write documents to facilitate decision-making by superiors and other stakeholders. This happens to be one of my strengths, and most architects have this strength.
I am a perfect combination of theory and practice. In my sophomore year, I passed the soft software examination as a senior programmer (software designer now). In 2014, I passed the software architect. The entrepreneurial experience made me realize the importance of channels; Guangzhou Dingfeng let me realize the importance of communication; In my last job, I learned that “less is more” : the chicken ribs (the needs that have little effect) were cut, and the late needs (the features that didn’t exist in the early stage but were useful in the later stage) were postponed.
The perfect combination of holistic view and technological accumulation nips risks in the bud. The same industry, the same company for a long time, easy to accumulate technology; Companies of different sizes and in different industries tend to have a big picture. Accumulation of technology first, and then cultivate the overall view, obviously better than the overall view after the appropriate technology. I have worked in large companies (Wuhan Kaimu, Southern Digital), medium-sized companies (Zhongshan Pole 3), and small companies (Guangzhou Dingfeng) for 4 years. Large and medium-sized companies can easily learn technology, while small companies (often without or with inadequate regulations) can easily find non-technical problems that cause project failures. Starting a business reveals almost all of its flaws (especially technical ones), and the subsequent years are a successful replenishment of relevant knowledge.
The common library directly uses the components I’ve accumulated to reduce costs, shorten time frames, and improve quality. In entrepreneurship and spare time encapsulation of a number of libraries (DLLS), three main: sn.DLL general function, does not rely on any library (including STL, MFC) and operating system. SNMFC. DLL, further encapsulation of MFC and Windows API. SNStl. DLL, further encapsulation of STL. At present, it has been verified by two commercial software and several small software. The source code can be shared between me and the company. Typical class: a few lines of code to set up server-side/client-side communication, but if you don’t use a class library, dozens of lines are enough to kill a novice. And automatic processing unpacking, sticky package.
CAD secondary development library. From 2017 to 2018, Guangdong XX packaged a CAD library when developing CAD drawing master, which belongs to Guangdong XX. Now I am looking for a job while repackaging the library. Compared with the library of Guangdong XX, it has a wide range of applications, but the performance is slightly worse. Typical class: An adapter class shields subtle differences between post-CAD2010 and pre-CAD2010 versions.
Develop high quality software Develop high quality software to improve the user experience. For example, software rarely crashes, maintainability (difficulty in adding or overchanging features, the biggest difference between architects and software engineers) is high.
Strong learning ability Strong learning ability to solve new problems. This is a result of adapting to the working environment, previous position: Common group software Engineer, Architect. Out of work needs: often complete all kinds of pre-research work. He served as the moderator of CSDN special topic development in March 2003 and the moderator of CSDN VC/MFC edition for many years in July 2003.
Training ability strong training ability, convenient team building. Online training 6 people, 4 people go to work (at least 2 people take high wages), 2 people (junior high school education) start a business to pick up work.
If possible, train one or more public programmers, code walkers, one at a time. Skilled public programmers and code checkers can greatly improve development efficiency: novice public programmers and code checkers can seriously reduce development efficiency: a defective public module means that all modules are defective. Novice code checker, will find a large number of minor defects, be checked in a dilemma: change, affect the construction period; Not to change, against the system.
I have the moral character to speak up and not to form cliques
Focusing on technology over business makes the difference between where I work, so I’m not afraid to speak up.
Practice is the sole criterion for testing a theory
No preference in scheme selection, according to the practice of adjustment. If someone else has a better plan, adopt it.
Only when enough is enough can one escape completely
It has something to do with my recent work experience. My salary is much higher than the average salary of a small or medium-sized company, which has no fault tolerance and if it screws up (or doesn’t work for 6 months), you’re out. If it’s done (or nearly done) and the company doesn’t have a job that pays me back, I have to leave.
Generally speaking, the technology in the early stage is important, while the technology in the later stage is relatively minor, which results in the output of the technology department in the later stage being lower than the income. At this time, I am willing to leave. In fact, many technical personnel, is also suitable for the early, they are willing to leave in the late, very depressed.
Good at summarizing
Summary of important content in written form for the convenience of those who come after.
This position is not suitable for technical managers who are trained in theory but are likely to be detached from reality (obsessed with quality, ignoring hours and costs). Even with the minimum CMM3, the ideal is 3 months, during which a lot of work is suspended. The team that runs fast, stops for 3 months, doesn’t know if they can run again. Even if it did, the speed would be significantly reduced, not conducive to market capture.
This task is not suitable for self-taught technical managers who are self-taught and have strong practical ability and are willing to experiment. But at high speeds, the costs of trying and failing are too high.