A total, B Total: Hello! I have been thinking about these days, our DOTNET team, what kind of road should be better in our company's product research and development. Here I really want to tell me some ideas. In fact, there are some words, I have already wanted to say, but I always have a concern, that is, I have to come, and the whole working time is not long, and the comments like this will be appropriate, will not Will be considered a lot, so I have been put in my heart without saying. But I now think that since I really intend to stay in this company for a long time, then you should say the idea in your heart. Come to the company for a few months, I think my situation is also more than the situation. Although I haven't been to the situation in other R & D groups, I still can understand some of them. The feeling for me is that the company's R & D department is not very similar to a development, but is like an ordinary business or productive department. Although I have been in a company, I only stay in a company, but from books, newspapers and networks, I think I still learned some other companies' way way. I think that as engaged in research and development, I should take the project-centered, and then fully mobilize the enthusiasm of employee enthusiasm in a variety of ways. Let me talk about the process of running the company's project. My original company implements a 40-hour working system a week, which is a theoretical provision, does not make such regulations, is not easy to recruit to employees, and it is not in line with the relevant provisions of the National Labor Law. To do this, it is the legitimate rights and interests of employees from the system, thus mobilizing the enthusiasm of employees. This is not the case in the actual operation. It may also be because of a small reason, the company adopts a typical project center system. Taking a six-month project as an example, the project will be divided into several stages, which is a number of milestones, six months may be about 5, 6 milestones, in the middle of each milestone, is also theoretically 40 The hour working system, that is, when doing progress, according to this standard, it is often not a weekend, or two days from the weekend, it may continue to work overtime every night, this is quite right. Flexible, determined by the project manager. If you work overtime at night, the company will provide dinner, however, after ten o'clock in the evening, the company reimburses the cost of going home to play a car. This is also the same as weekend overtime. After each milestone reaches, a certain amount of vacation will be arranged according to the situation of overtime, and the project knot will be arranged for a period of time. These vacation times are obviously less than the regular time. It is one to two days after the end of the milestone. After the entire project is over, it will arrange it before or after a weekend, so that the employee arranges his own vacation. In general, these measures will increase the enthusiasm of employees, improve the team's combat power, I turned out that the old pair said that it is better to hit people with fists. It is necessary to collect it to be more powerful. In terms of project operation, since I originally a Japanese-funded enterprise, I have a strong style of Japanese software companies, it is very attached to the design. In general, a project consumes the entire project during the design phase. The cycle is more than two or more even. The project design is detailed in the interface, that is, every CODER has a development task that has a hand in detail, and the CODER is only responsible for implementation. This is also a problem with the big business, and CODER doesn't know what you are doing. My original company is okay, the design is doing together, then the development task of the hand is still the purpose of this purpose There are two, one is to cultivate helping interns, and the other is to unify the style.
Another experience is to actively use component development. This is what I came to the original company, and all the components were all I wrote. We extract the features that are often used in the project to make the components alone, making the productivity greatly improve. Such as authentication, user identity, and permissions of the database, exception management, data-grade multilingual processing, page status save, etc. so that the development of the component makes it directly to reference directly to the project, and the efficiency is very high. Our company's AableCRM can only be developed in the month, which is the main reason. In particular, the page status saving component I developed, realizing unlimited page jumps, saving users, sorting information, according to some of some places in the 9CBS Forum, etc., there is currently no similar solutions. Ok, the above is my review on my previous experience. Here I want to talk about some of my ideas. I certainly don't evaluate the operation mode of other project groups of the company, I just said what I want my DOTNET project group how is it. 1. The development of the back project should not be as urgent like AableCRM, you should give me enough time to do design and documentation. I hope that the project later will walk into the right track. 2. In theory, a 40-hour working system should be implemented, five days a week, this is a basic premise of retaining employees. 3. Before the next project begins, as early as possible, it will recruit three programmers to enrich my group. Although I am still not very clear, I have some understanding, but I need it to five members. And a team with a tacit understanding is much more important than a few separate programmers, I need a period of time. 4. Take the project, after the start of the project, I will keep my team to maintain the highest efficiency work, achieve the goal as soon as possible, in order to achieve this, overtime is normal, but will be flexible by me, the milestone is reached After the project is completed, the vacation of a certain time is arranged. These things that are controlled by me should be filed with the superior. 5. The previous period is heavy, the specific gravity design in the project should be quite large, the design is detailed to the interface level, but the design of the back module can be carried out simultaneously with the encoding of the front module. 6. Pay attention to the enrichment of the component library, before and after the start of the project (especially after completion), there will be some time to analyze and summarize, put some good things, can be used in the next part of the project. In order to use it directly in the future. 7. Strengthen management, these these say that it is actually an accumulation of knowledge, we can't let this accumulation affect the company or project operation because of certain or certain personnel changes, using domain management and use of some collaborative tools are important. One point, code and documentation must be placed on the server. This work is put on the DS group, but in my DOTNET group must stick to it, I will not let my group develop free and casual style. Then I want to talk about a technology development direction of the company. As a uniqueness of our company's products, this is destined to have some products to run on the Windows platform, and I don't know other items, but I know some of the DS. Since our product is running on a Windows platform, then we should do some tracking in Microsoft's technical direction, using some Microsoft's new technologies, I know that a company is doing, technology transformation is painful, but facing Microsoft, face For the market, we have to experience some pain. But I don't know if I am willing to experience this pain. I am willing to learn something. I don't know how the company has been doing C . How many willing to turn to a platform that is still very strange, even if it is still in C , how much is the original The VC 6.0 upgrade is called 2003 VC 7.1.
But I have confidence, because since we choose the information field as its own development direction, you should be ready to learn new things at any time, eliminate the old things you are familiar with. It is possible to do this, there is some risk, but since I chose to compete in this market, how can I not face risks? If you are not willing to face the risk, how do you occupy the opportunity in the competition? These days, helped the DS groups made a criminal tool, which is a bit pain. This is a reason why I have not fully completed in these days. The result of the DS project is actually TXT documentation, honestly, this is in me. It is really incredible, it is really a bit ridiculous. The reason, one is that the traditional Win32 platform is difficult to analyze other data exchange formats such as XML support, and the other is not considered to be scalable, follow-up issues. This gives a follow-up development. For example, the development of my judgment tool has caused difficulties. I have to use regular expressions to analyze the entire document, the consequences of this: the operation is low, even though I have taken various ways To try to improve performance, but the regular expression itself determines how much performance cannot be improved; the analysis is not strict, this is a judging tool, although this tool will not be used to make a test, such as college entrance examination, but this after all It is still the exam, the result of the judgment is representative of certain significance, but you said that because natural uncertainty in plain text formation has caused my system to determine the score is not accurate enough, so the user thinks. In addition, I also heard other questions from the DS system, such as network transmission issues, buffer overflow issues, system compatibility issues, etc. So I started to consider the feasibility of developing DOTNET. It may be the reason why I have understood the DS system is not deep enough. At present, the possibility that the DS I know is that there are some operations such as network communication and DirectX. For network communication, the most direct SOKET programming, but significantly better practices are .NET Remoting, which makes the network programming of the DS is significantly efficient, safe and stable, .NET's name indicates that this is its strength . DirectX, DirectX 9 provides the Managed API, so that DOTNET develops such applications, which will be very convenient, and DOTNET's type security determines that the system's buffer error will become history, managed code also makes the operating system's compatibility problem It is possible to solve the price, the cost that is generated is that system performance may have certain, a small amount of decrease, but it is absolutely much higher than Java; some technical problems may be encountered during development, but it will be solved; developers This needs to be re-learned, this is the biggest price, to see if they are willing to learn. These ideas have been talked about these days, but I have to meet these, but I have to meet in the evening, I didn't talk about it. At that time, I said that I would like to wait until I will mention it later, such as the next one or the next few versions. I will consider it, but Ge always encourages me that I will mention it as soon as possible, so I will tell these ideas here. Another project I consider is a network language learning system that we will do in our DOTNET group. For this issue, the project has passed, and the project started to make a project of pure B / S. But I still feel that this is not suitable, I think I have been doing more than a year of B / S system, I confident that I understand B / S or more. B / S is not a million oil, there are many features B / S will never be implemented. If you look at the relevant reports of this time, we can see this very clearly, they call the C / S for rebirth, or It is a Phoenix Nirvana, Microsoft has to eliminate B / S, IE6 is its last version, and will not be further developed again.
Indeed, B / S is a lot of functions, but these are limited, many features are strictly subject to IE6.0, if not, the function will fail, the user disables JavaScript or installing the pop-up interceptor, the function will also be invalid Moreover, there are still many features that cannot be implemented. So I don't recommend using B / S, if it is necessary, I can develop a functionally simplified Web version, only some basic operational functions. These all uniform calls the system's WebServices interface. The above is my thoughts and requests, I hope that the company can carefully consider. Since the salute! Bill June 6, 2004
