Feasibility analysis of domestic operating system

Feasibility analysis of domestic operating system

Left light Hou
April 6, 1999

The Chinese need not engage in their own operating system. In my opinion, this question is not a problem at all. There is certainly a lot to discuss about "whether the Chinese have this technical ability" "to get something out of the market", but there is no room for debate on the question of "needing to have an operating system of your own". There are countless compelling reasons why we need to have our own operating system.
Of course, the need for light is far from enough. Under the existing conditions, can not make a practical, market-acceptable operating system to come.
The following is the author of the domestic operating system feasibility of some exploratory analysis. In the process of writing, I found that many of the views themselves are not quite sure (especially the first part, because I am not very familiar with the GNU things), if there is a fallacy, please give a lot of advice.
For the sake of simplicity, I call this envisioned operating system China OS.

1, the technology source of the OS

Many people, including me, have dreamed of rewriting Linux to become the Chinese operating system. Last year's "Computer" (36th "Linux Thought") once published in this regard: the use of C + + rewrite the Linux kernel and xwindows, to overcome the Linux to the common user too professional disadvantage, make it easy to install, easy to install and configure peripherals, Has a user interface that is as friendly as windows. With all the brilliance of Linux, rich application software, and worldwide technical support, this operating system may gradually get people's approval and eventually become an OS that can contend with Windows. The article finally called on the domestic software companies and the elite to jointly develop.

The idea is exciting and, in fairness, technically not very difficult. But in fact, it ignores an important issue. Linux is a free software that is approved by the gun general Public License (GPL). This license is guided by the idea of "guaranteeing your freedom to share and modify free software", which, while not prohibiting people from charging for the transfer of software and providing security, requires that the source code be provided at the time of the transfer, and that third parties must be allowed to use the license terms as a whole for free. It is extremely difficult to require a commercially available software to provide the user with source code, and the latter, which allows users to transfer the software free of charge under the GPL, has virtually eliminated the possibility of commercializing the software. This is exactly what the GPL is meant to do to prevent free software from being commercialized.

And our homegrown operating system--at least the one I imagined--should definitely be a commercial operating system. First, a free software like Linux can never pose a real threat to windows, because most of the market is not coincident, which Linus Torvalds himself admits. If the above operating system succeeds, at best can only become the pet of the computer enthusiast, occupy some less important user's hard disk. For security, technical support requirements, and so on, most important business users do not use a free software that exposes the source code. Second, even if we can convince our clients to use this operating system, we will face another serious consequence of free software. Because you can't collect money (it's useless to collect money, because users are free to copy the software to anyone who needs it), the operating system is simply not going to pay off, which means it has to really be owned by the public and not support any commercial company. Even if we are technically more than Windows, it is only "national technology", not "national industry". As long as we have a considerable scale of commercial strength, it is possible to say that China's national software industry in the world has a place. If China's operating system technology is the world's first, and Chinese companies in this area of the industrial value of 0, it is not a joke. And I don't think any company in the country will be paying for the project that doesn't pay for it--China seems to have no such big boldness of business.

In this scenario, the idea of rewriting the kernel to enable it to support Chinese and GUI with a friendly user interface is valuable, but there are a few ways we can get to the OS's technical sources. One is like Bill Gates of the year, starting from the bottom of the programming, to make a "fully domestic" operating system. Needless to say, everyone knows it's almost impossible. The other is like Linus, rewriting the source code of Minix, this is a relatively good way to go, but to make minix for commercial systems, also must obtain the consent of the copyright owner. I don't know if Prentice Hall will agree to that. In addition, there is a long way to go in order to use Minix for teaching purposes in harsh business environments. Another one is to start with the FreeBSD. As we all know, FreeBSD is a very good OS, providing the source code, its speed, stability and customization capabilities are not weak than Linux, the heavy network task support is better than Linux, and as far as I know, it is allowed to freely for commercial purposes. Its problems are more unfriendly than Linux, incompatible with hardware compatibility, and lack of Chinese support-which can be remedied by rewriting the source code.

2, the market positioning of the OS

This is the most important point. Here we compare the OS to other operating systems to analyze its strengths and weaknesses, and then locate the market it targets.

If the current Unix-like operating system can compete with Win9x in the field of desktop platforms, people with a bit of mind know that it is wishful thinking. Don't say that UNIX is Tian Shi with Win9x in ease of use and interface, even if it is repackaged and has a simple installer and friendly interface with Win9x (or even better), it is not possible for ordinary users to accept it anytime soon. Because people have become accustomed to operating under the Win9x environment, and have become familiar with the various applications under Win9x, it is very difficult to turn to another completely unfamiliar environment. Let's remember that famous example: Someone who has a new keyboard that has improved the key layout and is 30% more efficient than the traditional keyboard has failed in the market because people have given in to their laziness and are unwilling to learn new fingering. What's more, the strength of the OS (if it does)--cheap, stable, and secure--doesn't appeal to ordinary users who are not sensitive to these aspects. At least for the time being, the best market position for the OS is to settle on a server rather than an ordinary desktop.

So what is the feature of the OS compared to popular server operating systems such as Solaris, HP-UX, Windows NT, and so on.

Win NT's advantage lies in: with Win9x compatible operation method, and the complete culture of the interface, make it in the ease of use, any one familiar with Win9x can very smoothly grasp the basic application of Win NT. If you have a certain network knowledge, use it as a server operating system can say that hardly need to spend much effort. With the development of China's network, the network managers are relatively insufficient and the level is relatively low, Win NT for other types of operating systems obviously have greater temptation.

However, the Chinese network must never mature to mature. In the near future, when the demand for the network is increasing, the level of network managers is increasing, Win NT will gradually reveal its weakness: poor stability, poor security mechanism, poor openness ... At the same time, with the gradual regulation of the copyright market, the high price of Win NT will be a part of users. If we can launch the Chinese OS based on GUI and kernel support, as mentioned above, all these drawbacks can be turned into advantages. Not only does the OS have an advantage in stability and price, but as a member of the large UNIX family, its support for the network is inherently superior to win NT. Although Win NT is also making progress, it is inconceivable that it will surpass Unix-like operating systems in the areas listed in the near future. What's more, China OS as the Chinese own source code of the operating system, in the climate, geography, people and all aspects than is trying to "localized" Microsoft to have the innate advantage.

So what's the advantage of the big business Unix system represented by Solaris?

The first of course is the price, the OS can have a much lower price. You might say: "For a large commercial project, this factor is meaningless." "I admit that the first phase of the OS is not enough to confront the above OS in the high-end market, but we can try to avoid that in the market position." Like a typical intranet, a large number of Web servers, small and medium sized database applications based on C/s, and so on, people will not install a Solaris-style jumbo for each network. In fact, this is also a win NT positioning of the market space, but I think in this market, the home OS on Win NT has a full advantage.

To sum up, the domestic small and medium-sized server market as the main target of China OS, in the competition to occupy a place is very possible. This is a very large market, as long as the share of a certain size, it is enough to maintain a large company. With a firm footing, we can guarantee that we will not fall in the short term, and that there are many opportunities to move to other markets in a step-by-step manner. Although this assumption is not without risk, but the old saying well, there is nothing in the world is completely risk-free, not to mention the high risk in the IT field. And once the idea succeeds, there is an infinite business opportunity behind it.

3, the technical requirements of the OS

(1) Rewriting of the kernel

Fully rewrite a UNIX system so that it meets the following requirements:

In the kernel on the basis of international standards support Double-byte, is a thorough Chinese OS. Provides two Chinese support environments in GUI and command line mode. Provide all the Chinese documents.

It can be run without modification at least most Unix software.

Provides a complete GUI interface like Win NT, simplifying operations to the fullest. From the present situation of Xwindows, one step still has a long way to go. In fact, using a GUI to completely replace the shell is not even possible at the moment. But we don't seem to need to use the GUI to do all the functions. I admit that a full GUI loses some of the features of Unix, but the conversion of DOS to Windows is a successful precedent.

Given its market positioning, the network part of its core must be extremely stable. (Linux is passable in this respect) software that supports the network is off-the-shelf, and Apache and SendMail are superior to IIS in performance, which is known to all. We can make them completely Chinese--in this respect, using free software doesn't matter.

Supports the latest hardware and simplifies hardware installation and setup programs.

There is no need for a powerful office suite, as it will be used primarily for servers. But we do need a more powerful word processing software than VI or even xemacs, just as we need more powerful ultraedit than Notepad.

(2) Database system

From its market positioning, the OS must have a strong database system. There are two solutions: first, to take copycat, through the cooperation of the way, the existing database system directly transplanted, and to be Chinese. Large database systems such as Oracle, Informix have unix-based versions, and their providers are aggressively planning to enter the Chinese market, and they have a common enemy, Microsoft. The second is to develop their own hands, which is more in line with our intention to develop the OS. Of course, the development of such a system, the difficulty is not under the development of an OS, and we do not have freesql to use. But in fact, some people have tried this in China, such as the Dream Database company introduced in the 12th issue of computer Business intelligence. China's software power in this area to achieve what level, I do not know, but whichever way to choose, ultimately will be to build our domestic database system software.

(3) Development of language platform

It is not enough to have a database system and some applications inherited from UNIX, and if you want to become a popular and competitive operating system, you must provide an excellent development language platform. Because the OS is a complete GUI for class windows, a GUI-based language development platform should be designed and provided with a complete set of APIs like the WINAPI function library for programmers to use. Perhaps, it can be divided into two levels: one for advanced developers of the lower-level language (like VC), one for the less demanding software developers of the RAD language (like VB). Of course, all of these must have a full Chinese version.

With such a development language, you can develop an extensive two-time development of the OS, possibly with Office, browser, engineering graphics software, and image processing software on the OS. Of course there are games.

Digression:
The idea was so large that I wrote it later and felt a little unrealistic. In fact, this is almost already building another windows. But I think and think, or take the liberty to open it, hoping to play a role.

One thing is for all to see: Microsoft's offensive in China has become more ferocious after the market for desktop users has been breached. Windows NT and SQL Server have become more and more server solutions. ASPs are also increasingly becoming popular web scripting languages (and, of course, the reason for their own advantages). I do not dare to comment on the trends in the market, just feel that we do face a danger: not only the monopoly of Microsoft, but also to a considerable extent excluded from mainstream computer technology, because in foreign countries, Unix still dominates the world trend. Of course, maybe Win NT will eventually defeat Unix (a little weird.) Or UNIX vendors will occupy the Chinese market as much as the mainstream of world markets. But to step back, Microsoft's monopoly is bad, Sun's monopoly is good. The pace of the IT industry is getting faster, and new and bigger historical opportunities are in sight. Maybe it's time for the Chinese to think about something.