Prepare to write a software rendering engine.

In last June, I became interested in 3d graphics. I have read a lot of books, and I have never had time to write this code because of my busy work. This time I am going to organize my understanding of graphics. Although I have read "3d game programming master skills" several times, after all, it is not as profound as I did. Writing through my own thoughts will also have a sense of accomplishment. It has been almost a month since I completed the compilation system, and I have read some books that I have already read several times. This is certainly faster. I read the c ++ books of com and scott meyers again. Speaking of scott, after hearing his lessons, he thought he was really a world-class master. Not only did he have a high level of skills, but he was also very clear about his presentation, and his pronunciation was also very standard. Although he was a little quick, however, I listened very seriously at the time. Even c ++ had a better understanding after that time. I also looked at the design mode. I started reading the design model in the third year, but I didn't understand it at that time. Although <design pattern> the book was indeed well written, it was not suitable for me at that time. By now, the design pattern analysis was quite suitable for me, so after n times (n> = 3) of reading, I finally got started. I usually write code on my own and I also noticed some patterns. Later, I checked out this "java and mode" and combined it with the fishgui project in "tao nature" to learn the design mode from the specific code. Later, several models were used in actual project implementation. But now I think that the design model is nothing more than a summary of some good designs by our predecessors. What truly solves the problem is our own ability to analyze the problem. It is not worth spending too much time on this formal thing. The essence of oo should be encapsulation changes. All design patterns seem to be able to do this. This reminds me of uml. uml is only a way to express developer's design ideas and analyze and solve problems. Even if you don't understand uml, you can use a pen to draw some diagrams on paper. Now everyone is aware of these things and will not pursue them too much. I am no longer so interested in this aspect.

This time, I plan to use the lower design mode and compilation to deepen my impression and may use the lower com. I still want to figure it out when I see it. After all, to turn c ++ into com, I just need to encapsulate it in multiple layers and write some idl and add the interface. The mesa (open-source opengl implementation) is coming down. Now we can use vc to tune it and run the gear demo, the implementation with gdi can also reach 18-20 FPS, good. Even read its code.