Theoretically, as long as the corresponding decoder is installed in the operating system, the corresponding media can be supported.

The fifth chapter is a post-reading experience. In general, the viewpoint is consistent. In particular, to design a new project, solving the core technical problems is the key design method for success or failure, I think this is only the first step of the long journey (of course the most critical step). There are still many things to follow, such as project planning, system architecture, and design of various modules, ignore every step. In fact, in an open network environment, as long as new technologies are not required, it is not difficult to solve problems with mature technologies.
In addition, he gave some comments on the hero's software. Lao Liang said that the solution system was designed with "open thinking" and various new functions could easily be added. But how can we solve this problem? Why don't I use a single player to play all the media? What should I do? In addition, the business grade of the software is not enough, and it still feels like "handmade", especially those tools, just like a semi-finished product testing product. Hero seems to want to do everything by himself. The installation interface, menu, and dialog box are not very standard. Now commercial software operations are getting more and more similar, and users are easy to learn. In addition, if the architecture of the solution is open, how can we not automatically download the module to install new decoder, patch, and other automatic update software? (I bought two patches in less than a month for the solution 2002.) Check realplay. mediaplay automatically processes everything and the decoder can be downloaded at will, theoretically, only the corresponding decoder is installed in the operating system to support any media. The book also states that Windows is an open system, Linux is a closed system, and Linux is a kernel. I don't know if Linux can use any plug-in module? Only the most important modules are compiled into the kernel based on hardware conditions. Other modules can be plug-ins and can be adjusted at any time. This is an open system and win won't work, no one can change the kernel. In addition, the win api is too huge. Everything is provided to the API, and all the MS packages are provided. You can only call this API. In addition, it seems that if you want to add a CPU or something in Win2000, You have to reinstall the system. In Linux, you only need to restart the system to automatically identify the hot swapping of the CPU, and you don't even need to restart the system. What's the problem? Windows seems to be a GUI or GDI graphic interface, but it seems that it is also the first choice for UNIX to play games and make movie CG special effects, not to mention the server.
In addition, especially for those code specifications caught in the book, I thought that a "{}" indentation of a "tab" was a typographical error. I did not expect it to be a standard! Who will look comfortable with this specification. In addition, in coding, for the purpose of unification, "tab" is replaced by "insert Space" and there are four spaces. In this way, the editor is more comfortable and will not be messy. These Regulations do not know whether the hero is enforced, or who will be willing to work?
The COM interface is not accurate enough. The user calls the COM interface to point to a actually com stub, and the COM component may be in another process, it may also be in another machine. Stub may pass the parameters and data objects of the called interface through inter-process communication or even the TCP/IP protocol. Com is not a simple DLL. dll is just a carrier of it. EXE can also be used as its carrier.
However, Lao Liang has a good explanation of Hook Technology and DLL usage, which is very thorough and open-minded! I just want to read this, understand it clearly, and then read "Win95 system Mysteries" and "core programming" in detail.
Thank