First, the project should be managed as a whole, from the beginning to the end? At the beginning of the whole project, the overall plan of the project should be made. Throughout the project, it should always be implemented according to the project plan. If there is any change in the project, an impact analysis should be made. After approval, a change plan shall be made and implemented according to the change plan. The impact of the change, such as cost, schedule, etc. , the relevant project stakeholders shall be notified to explain the reasons and impacts of the change. ? The start and end of the project is also an important work in project management. It is necessary to sort out and file the documents and materials produced during the project; Review the project cost and progress, and check and accept the project quality; Summarize and communicate the advantages and disadvantages of the whole project process. ? Change plans are often encountered in software projects. In order to control the change of software project, it is necessary to determine the initial target benchmark of the project, and the user needs of the benchmark are clear, so as to measure what needs to be changed. Otherwise, it is impossible to make a change plan and delimit the boundaries of change if the changed things are mixed with the things originally needed. ? In order to occupy the market and win the contract as soon as possible, I signed a contract with the user before providing the user requirements in detail. Later, not only the cost was limited, but also the time was not enough. During the project, users always change the functions and requirements of the software. Because there is no basis, we think it is to change the demand and add new functions, while users think it is the scope of the contract, so we can't increase the cost and time. At the beginning of this project, it seemed that we had signed the contract and won the initiative. In fact, the demand is not clear, which makes us very passive in the later project process. ? Therefore, it is necessary to make a good plan at the beginning of the project and make clear the objectives. Only by defining the objectives of the project, reasonably arranging resources such as time, cost and manpower, and controlling the changes of the project are the basic conditions to ensure the smooth completion of the project. ? Second, the project scope management theory solves the problem of unclear initial demand of the project? Requirements management is a problem in project scope management, because it is actually the premise of all management principles in the development process. Only when the goal of development is made clear and understood can software development be carried out in a planned and orderly manner. In fact, if there is no documented requirement, the product is likely to deviate from the requirement before and after the development work is completed. Planning, tracking, configuration management and software quality assurance involved in other key processes all start from a stable foundation, that is, a documented requirement baseline. ? What needs? Requirements refer to "system requirements assigned to software" or, more succinctly, "distribution requirements". These requirements may be technical (such as functional and performance requirements) or non-technical (such as release date and expenditure restrictions). ? It is important to distinguish requirements management from software requirements analysis. Once the release requirements are documented and passed by all affected departments (customers, system engineering, software engineering), the basic work of requirements management is completed, and the rest is to manage changes. There is no evidence that the distribution requirements themselves can be used as the whole foundation of software development very clearly and completely. In fact, usually they are not. ? It is the task of software requirements analysis to optimize and accurately describe requirements, fill loopholes and express meanings more clearly. The result of analysis is called "software requirements". In this way, distributed requirements as the output of requirements management actually become the input of software requirements analysis. Requirements management is far ahead of the technical action of software development, and software requirements analysis is the first step of key development technical action. ? From the description here, the activities of demand management seem to be too simple and too basic. Obviously, no software development organization can effectively carry out such activities. The problem often lies in the fear of transparency. Customers feel that keeping the requirements vague, loose or without formal documents can give them more opportunities to say, "That's not what I want, that's not what I think the requirements mean." The requirement of clear documentation may force users to be responsible for starting changes when the system meets the documentation requirements but does not meet the actual requirements. Similarly, developers feel that vague, loose or undocumented requirements can give them more space to get as close as possible to the budget and schedule, and then say, "This is what we think requirements mean. If you need something else, you have to pay another price. " Documented and clear requirements will force developers to undertake the obligation to meet these requirements, and expose them to the risk of inaccurate expenditure and progress evaluation. ? In this way, although the interests of customers and developers are opposite, they have come together. Each party thinks that it is protecting its own interests and consolidating its negotiating position, but in fact, each party is disappointed and quarreling in the future, laying a time bomb for the project. ? Third, how does the project time management theory guide us to grasp the main contradictions in project management? In the past, project management was based on experience and everyone's work characteristics. Software projects were basically carried out according to several main processes, such as requirements analysis, general design, detailed design, code programming, debugging and testing, and user acceptance. But the project division of labor is more detailed, how much time is expected for each small process, how much time can the whole project be completed in the shortest time, how to arrange personnel reasonably, and how to grasp the key links in the project. , has not been quantitative analysis and management. ? The most intuitive implementation of project management is to shorten the project time. There are many examples of shortening time by using the theory and method of project management. During the overhaul of Louisville chemical plant in the United States, the overhaul process was finely decomposed according to the guidance diagram and the control relationship was established. They were surprised to find that the total time for choosing different paths in the maintenance process is different. By repeatedly compressing the task on the longest path and repeatedly optimizing the construction period, the usual maintenance of 125 hours was finally completed within 78 hours, saving 38% time. This is the famous time management technology CPM, that is, "critical path method", which is still applied by project management workers. ? Therefore, in software project management, we should also apply time control theory, combine the reality of software engineering, decompose the tasks in more detail, establish the whole workflow with network diagram, estimate the duration of each stage, find out the critical path, shorten the time limit for a project of the critical path through quick follow-up and improve work efficiency. ? Fourth, quality management is the key to the success or failure of the project? In the process of software project, we always think that the functional testing of software is serious and strict, and every test must have a test plan and use cases before testing; The test shall be recorded, and the records shall be compiled into a test report. ? However, after passing this training, I feel that our testing work is still far from the requirements of quality management. Quality control should go deep into everyone related to the project and every process of the project. From the beginning, we should establish the concept of quality first, and ensure the quality of each process, instead of thinking about quality only when we have the best testing, and then measuring whether it meets the standards. ? Standardized design and standardized management are the guarantee of engineering quality. Participating in quality system certification is helpful for enterprises to improve project management level and project quality. The CMM model has been widely recognized and accepted, and CMMI follows its organizational model, with five levels and 18 elements. Through five-level certification and strengthened management, enterprise project management can be upgraded from chaos to milestone inspection, to clear management system and standard, to quantitative management of statistical process control, to final optimization of process, evaluation of workflow and improvement of workflow. ? I have participated in some functional design and programming for Japanese software before. Japanese software companies have carefully controlled the quality of a project, which is simply intolerable to our concept. The detailed design of a module should be described by the standard graphic language provided by them and explained by the standard design template; After the design is completed, organize relevant personnel to evaluate the design. If there is a problem, the design should be modified, and then coded as a design document until the evaluation is passed. The code is finished, but the result is not finished. In order to print out the code, the relevant personnel evaluate the whole implementation process of the code and put forward suggestions for modification. After the code is modified, it needs to be retried and can only be submitted to the code base for code assembly later. ? At that time, it was thought that the Japanese method was a waste of time and manpower, and the ability of technicians was too low. How to improve work efficiency? However, the frequent occurrence of software quality problems is that we constantly realize that starting to waste some time and manpower and controlling the quality of every detail is to save a lot of new time and manpower expenditure for solving quality problems. Save a lot of software quality maintenance costs in the later stage. Generally speaking, it is accounting. In order to improve the project quality and reduce the cost, quality control must be done well from the beginning of the project. ? 5. Can the application of some strategies in communication management make the project better completed? Doing a project requires contact with customers and there will be some formal and informal negotiations. Both sides will bargain for their own interests. Good communication with customers is the condition for the smooth progress of the project. There are many strategies in communication that can be used in daily work. The purpose is not to harm people, but to better complete the project and achieve the goals set by both parties in advance. Communication skills include: issuing deadlines, using surprise attack method, using limited rights method, people who don't show up, being fair and reasonable, delaying tactics, reasoning together, retreating, irrationality, fait accompli and so on. I successfully adopted the strategic delay method to urge the customer to recover the project quality deposit in time. ? There is still a lot of experience. In short, through this study, I have a new understanding of project management, and I will apply these theoretical knowledge to practical work. In order to improve the project management level, improve the project quality, reduce the project cost, reduce the project risk, and finally improve the enterprise benefit.