Technology in the eyes of an entrepreneur

(I am not a good man, and I don't have much time. I can't write any brilliant opening remarks. I can only write anything I want to think of. I hope you can forgive me !)

 

Background:

Five years after graduation, I have been engaged in C ++ development for five years. In the middle of last year, I decided to embark on the road of entrepreneurship. Now, I have had a lot of white hair on my head.

In the past year, you were either CEO, CTO, or Cho. You may have guessed my situation...

This article will not talk about other experiences, so let's talk about technology :)

 

First, we should discuss the following: technology is not a problem!

Many technical staff not only like to talk about this sentence, but also think from the bottom of my heart that I used to belong to such a technician. After the following two stories, I woke up completely.

One is the first project I received when I started my business. The demand seems very simple. After a simple analysis, I felt that there was no problem with the technical implementation, so I started to writeCodeAfter the test is completed, it completely does not meet the expectation, and the problem lies in one of the seemingly simple sub-requirements: data packets are forwarded between the two network ports on the insert board. I think you may despise me a little after reading this demand. what technology is involved in such a simple requirement? Ha, don't worry. I will add a condition for you: the packet is forwarded between the two network ports on the embedded board. The delay of each packet forwarding is no greater than 10 us. The first half is provided by the customer, and the last half is analyzed. Of course, the two are indispensable. In the face of the latter half of the sentence demand, I am stupid, really. At that time, the operating system was Linux, and the previous implementation of the application layer must be fully charged, because the process scheduling is at the MS level, and the US-level shadow cannot be seen, my friend suggested implementing it in the kernel. Wow, the kernel, I used to work on the application layer stuff in the last five years. I have never played with such advanced things as the kernel, but for projects, for the sake of money, I have to stick to my head. Fortunately, with the help of powerful Internet and enthusiastic technical experts, I finally know how to write code at the kernel level, I also know how to work with the NIC Driver. After the code is completed and tested, the progress is not small. The forwarding speed of most data packets is reduced from several hundred ms to several hundred us immediately, but it still cannot meet the requirements. First, almost all data packets must meet the conditions. The "majority" of data packets cannot be used. Second, there is still a lot of distance from several hundred US to 10 us. What should we do? If you try a real-time patch, you may be able to improve the situation. After the patch is completed, the "majority" is changed to "Almost all ", but hundreds of us remain unchanged. At this time, I thought of changing the real-time operating system for development, but the board has already been selected by the customer, which involves a lot of commercial interests. It cannot be changed, and the Board supplier has no experience in real-time operating systems, it's impossible to do it for you. Without their support, you can just take a break. Finally, some people suggest looking for a solution on the NIC hardware, but they can find it. There is no fully qualified Nic, and the vendor has to design it for us, but his mother had to pay a deposit of several hundred thousand, and this was passed again. Maybe I can find a hardware engineer to help me design the project, but there are no such people around me, and this project has been dragging on for a long time. The customer has no patience and can only give up. If anyone who sees this blog post can solve this problem (any board can be used), contact me to make money together.

 

The other is the ERP project, which helped an enterprise develop a warehouse management system. At that time, I also found a friend who worked in a well-known ERP Enterprise to help me understand the needs in the early stage. It is good to have experts help me, the customer's requirements were all in my friend's expectation. One afternoon I figured out that no matter the requirement or technology was difficult, so I took over the work later. From the beginning of code writing to the first measurable version, it was still very fast. During this period, the customer company also provided an interface person to communicate with me and thought it should be okay, once the product reaches the customer's hand, the number of people asking for the product suddenly increases, and the contact also says that there is no way. At that time, he asked other people and said "OK". Now they don't accept it, he didn't offer it either. He could only ask me to change it. Otherwise, he couldn't get the money. Okay, I changed it. When the change was over, they looked at it and thought it out again, every change is a little bit, but a little bit of it can make people crash. In the middle, I am a patient, while changing, they change, I spent half a year on a three or four-month project, and I was overwhelmed.

 

The first one is purely technical, and the second one needs to be mixed. It is a problem for me. Since then, the technology is not a problem in my mind.

 

Second, technology is only a tool in your hands. It is a tool that may be discarded by you, but your thoughts can be left behind.

I have been engaged in C ++ for five years and have no deep feelings for C ++. But when I was developing this project, I had to abandon it. It was very simple, I cannot afford it because it is too expensive. What do you mean, expensive or expensive? Ha, if you are already a C ++ company that has been working hard for a few yearsProgramClerk, you should know how spam the C ++ code you wrote when you first debuted. After a few years, you know a lot and your value suddenly soar. Happy! Of course, I want to congratulate you on this, because you have survived. But as the person in charge of a newly-founded company, I must consider the money in my pocket. First, I dare not use new people who just debuted. c ++ code should be well written and it will take a lot of training costs, after training, he ran away, so I don't want to cry. Second, the price for finding experience is high. I can't afford it, so I just look at the face of money, I had to give up on it. Of course, there are more reasons. Before talking about other reasons, let's first talk about who I chose, Java, right, Java, many people, cheap, easy to use, and many free resources, in addition, I want to develop a B/S system. There are many mature Java frameworks. Unlike C ++, I cannot find any well-known frameworks (not to mention ace, I can understand this guy's cpper is not much, and it is even more rare to be able to play well), so I chose to select only java. As for the client, I used flex, both development efficiency and user experience are good. As I said, C ++ seems useless! From getting started with Java and flex, to using them to write code for a week or two, the internal reason is that they and C ++ are both object-oriented languages, they all inherit the idea of object-oriented unification, and use one of them. Other brothers' languages are also easy to understand.

 

The ass decides the head. This sentence is really good.

 

Finally, to gain an in-depth understanding of the ecological environment of your industry, not only to find a direction for the development of technology, but also to find a direction for your life.