The concept of game development is changing again-need to learn more

Come to the West Hill to work for some time. Like me this small company came over, in addition to the canteen of the food is more delicious, also did not find that big companies have what special place. 9.25 We began to have a holiday, a lot of function orders as of the date is 9.25. I've been in a hurry this past week. Because the system is about to re-do it, the time is quite hurried. But today the discovery function must have a single finish. Since it is not finished, there is no way. Let's get down to the chase.
This week I was debugging my buddy system. This system I took over, the first time I took over, said the server has been completed, access to the client's new UI. It happened that I started to do this function, and the planning side was urging. I probably looked at the service-side code, I feel like overtime, I was in Friday received a single, I thought this Saturday and Sunday are overtime, the next Wednesday should be able to finish, and planning to say the next Wednesday fix, and then engage in the next content. The next step is to work overtime with bitter force. By the Monday, the planning temporarily under a slip to wait, only to let go of the hand. After finishing, the next day was Tuesday, the boss said to give a service-side design. I thought the code is almost, directly to the right, and then found some problems, to change, since it is a problem, then I changed. The next day began to rush to, my solution, I thought the code is almost finished, one can be right. The boss then walked to my card and asked me why I had to write the code. I was saying "finish writing the code to determine the final result". Later I think about it, I actually want to express the meaning of the game in this play, in fact, there is no need to spend time to design what it, in general, the storage of data to think well, began to write code, after all, progress is the first. If there is really any problem, I think it will be quick to come over, up to half a day to finish. After all, it's just a UI play thing. Because before the page tour company in Guangzhou, the competition is really very intense, the rhythm is very fast, so the function comes down as soon as possible to complete the best, there are problems also rushed to solve, this way. So the way of working here is somewhat different from the West Hill. And before the need for code review, now in this project group without code review, that is to say that the pre-completion of the code can be done, and we are quickly completed, the other team will review the code, so although the speed, but the general problem is not too much. There is also a point is the design design, I have always felt that there is no content to write, are playing the same logic, write code is good. Now I know, first draw a good flowchart, and then paste the key data structure of the design can be. Now that I know what the design is going to say, I'll finish this part of the work next time.
Another point is back-end development. My previous company A dose, about tens of thousands of people. Less, the back-end logic will not be too much of a problem even if it is inefficient. As long as the bug is not good. So my view at the time was that logic was simple and clear and inefficient. Now do hand-travel more feel not to focus on the implementation of the efficiency of things. So took over the mail system, found a query mail operation distance has three query database, I then and the previous developer confirmed, and later felt no problem. But there is a problem with the pressure measurement, the mail efficiency is very low. Later, I discussed the transformation of the mail system with my colleagues. Listen to another colleague who said that one of their previous projects was stuck in the mail system for the last few days, and I realized that it was time to focus on efficiency and learn more about MySQL.
Generally speaking:
1) offline player data as little as possible to load into memory, to consider the medium-term registration of the players a lot but many are not used, then the full load of data into memory is very bad.
2) MySQL still has to pay attention to the efficiency. A table player, for example, saves one row of records. Added 10w players, each player 5 records, there are 50w rows. If the index is a string, it will be less efficient. So try to save one row of data for a player, and multiple records to be serialized into binary and then saved to a column with something like Protobuf. Keep in mind that the records in the MySQL table have always been inflated and only increased.
3) There is a need to seriously consider whether there is a possibility of inefficiency in MySQL when the player needs to log in.
4) Serialization is really a great thing to be useful in json,base64 coding projects.
5) The stress test of a large company is a standard process that provides data to indicate which functions are inefficient in the project, which is still a great thing.

The concept of game development is changing again-need to learn more