[Cultivation 4] technical advance research

In June 19, when pre-research items are included in the plan, we will conduct pre-research work. The work at this stage will tackle major risks in the project or the uncertainty of the technical solution. It is an important challenge for the project manager to efficiently complete this stage. I have seen many project managers have a good grasp of the overall management of the project, but the project is out of control due to insufficient grasp of the work in the pre-study phase.
The five-factor decomposition of the pre-research phase is as follows:

1. Stakeholder ExpectationsStakeholders at this stage are pre-research personnel and product managers. 1) the complexity of the pre-research plan should be weighed by the input-output ratio of all projects within the planning expectation. Before the pre-research starts, we need to know the psychological expectation of the Supervisor/product manager, if the implementation complexity of the pre-research solution is far greater than expected, it should be proposed in a timely manner and discussed. If necessary, the pre-research should be stopped. 2) The main purpose of pre-research to eliminate implementation risks in the pre-research phase is to eliminate the implementation risks of the entire project. At the end of the pre-research phase, there is a consensus that the project can be implemented, the workload estimation of the entire project plan will also be based on the pre-research conclusions.
2. SandboxTo do a good job in technology pre-research, there must be two prerequisites: a clear target value and an operable acceptance method. 1) There is a clear goal, for example, to pre-develop a solution. The final submitted solution requires 1 GB of memory, and this memory consumption cannot be accepted. In this case, rework will occur, at the same time, pre-study engineers will complain about it. Why don't they say it too early? It hurts and hurts. Therefore, the general method here is to issue a pre-study target book before the pre-study starts. The target book contains detailed restrictions on the resources consumed by this pre-study, such as memory, CPU, and stability.
2) if there is an operable Acceptance Method and there is no operable acceptance method, the target will not be executed and will become empty-spoken or out-of-concept. Therefore, it is important to confirm the acceptance method before the pre-study starts. Example 1: objective: the new pre-research method has no more than 5% impact on the performance of the original system. acceptable methods: load this module or not load this module, count the consumed CPU cycle in three identical business flows to determine whether the CPU consumption exceeds 5%.
Example 2: objective: the response speed of a complex front-end page is returned within 3 s during normal use by the customer. Acceptable Method: test the page response speed without CPU load under two typical customer business flows.
3. Plan Formulation1) The overall plan of the pre-study project is usually 100% accurate, especially for the customization with high risks. However, the reference time of an overall plan is as follows, will give people a goal to get a sense of direction.
2) quick follow-up is the most reliable way to follow up the plan in the pre-research phase. It is difficult to have an accurate plan when the plan is not confirmed, but we can adjust the current plan-rolling plan in time through quick follow-up. Two common methods for fast Follow-up: phase and time. Phase: the current pre-study is divided into several stages (for example, environment setup, train of thought 1 attempt, and train of thought 2 attempt). Each stage reports and discusses the work situation. Time: the pre-study progress is confirmed every two days and every one week. Phased follow-up can be adopted for simple and well-defined pre-research, and time-based follow-up is required for complicated and difficult pre-research. Of course, you can adjust the mix of the two methods according to the project requirements.
4. RisksTypical three cases: 1) Excessive pre-study, common in the case of prudence or lack of technical confidence in the project manager. They need to eliminate all risks in the pre-research phase. Because the code in the pre-research phase has not passed the design phase, it is difficult to reuse the code in the subsequent project coding process, resulting in a waste of work, as a result, the entire project cycle becomes tense. 2) lack of pre-research, which is common in project managers who lack big picture or strong technical confidence. They often select several risks for important follow-up based on their own judgment, they often inadvertently ignore the risks of areas they are not good at or unfamiliar with. As a result, after the project enters the subsequent stage, disasters occur frequently and fire fighting is always in full swing. 3) The pre-research idea is incorrect, leading to heavy rework.
The method to eliminate the three typical cases is to form a sound technical decision-making chain as early as possible in the project team. Here, we need to rely on team management to provide the foundation.
5. Team Management1) The core personnel of the project at this stage should have been or will be in place one after another. At this time, the project kickoff meeting should be started. There are two main things to do: A. Describe the goals and significance of the entire project. Anyone wants to do important things on their own. Since the company or department decides to invest manpower to start the project, naturally, there is an important significance. At this time, we need to convey this significance to the project team members so as to mobilize everyone's enthusiasm for work. As John Scott said: "If you want to build a ship, what you need to do first is not to collect wood, process boards, and dispatch work, but to arouse people's yearning for the vast sea. "B. centralized Office-move the seats together. Most of the time, the project team members come from one or more departments. The previous seats were not tied together. After the project team was founded, we should move the seats together, centralized office is one of the most efficient ways of communication.
C. Explain the plan to the pre-study personnel to understand the significance of the pre-study project. The purpose is to enable the pre-study personnel to understand the purpose of starting the pre-study, generally, when we study something in depth, it is very likely to trigger other ideas to solve the problem. The clear purpose can make the pre-study staff flexible in some cases. For example, the project team planned to provide Xiao Wu with the goal of verifying whether the MSI multi-queue of the NIC can be implemented on the linux2.4 kernel. Objective: To improve the performance Xiao Wu finally gave the pre-study conclusion, it takes a lot of effort to port Nic multi-queue to 2.4, but it is difficult. However, some MSI functions are transplanted and the trigger method is changed to the timer, this improves performance and reduces the workload.
2) It is necessary to clarify how to follow up the technical solution in the pre-research phase of the technical decision-making mode or how to follow up and handle the difficult problems. The benefits are as follows: A. When the project team member gets stuck, they know who to ask questions B. If the project manager is not responsible for technical issues, the technical decision makers can clarify their responsibilities.