The 1980s Chinese OSes That Beat MS-DOS at Home—and the Ecosystem That Killed Them

We had domestic operating systems as early as the 1980s, but why did Windows ultimately dominate? The answer lies buried in time.

Recently I had some free time, so I took the opportunity to study Fastify and HarmonyOS development. While learning about HarmonyOS, I unexpectedly came across some very interesting "history"—the history of domestic Chinese operating systems.

Beginning

The story of Chinese operating systems starts in the 1970s. Up to and during that decade, China was unquestionably "poor and blank" in computer development.

Domestic 103 computer (image from the internet, contact to delete if infringing)

The domestic 103 and 104 machines were inventions of this period; at that time there wasn't yet a true operating system.

It wasn't until the second generation of transistor computers that the "shadow" of an operating system appeared, then called "supervisor programs." In this phase, China produced DJS-220, DJS-150, and DJS-240, which can be considered the pioneers of Chinese operating systems.

In 1981, Microsoft officially released MS-DOS 1.0—to be precise, Bill Gates bought it for $50,000.

MS-DOS 1.0 operating system (image from the internet, contact to delete if infringing)

Back then, IBM commissioned Microsoft to develop an operating system. Gates "married off" the purchased 86-DOS twice: it was called PC-DOS on IBM machines and MS-DOS when sold by Microsoft. (I wonder whether IBM signed an exclusive contract.)

MS-DOS was the first 16-bit operating system to see real-world deployment, and during the 1980s it became the standard OS for PCs and compatibles.

In 1983, China's first operating system for a supercomputer, YHOS, was born with the Galaxy-I supercomputer. In the same year, a star-like operating system appeared on the domestic market: CCDOS.

CCDOS V4.0 original floppy disk and user manual (image from the internet, contact to delete if infringing)

CCDOS solved a very critical problem: Chinese character storage and display.

By 1987, Microsoft released MS-DOS 3.3, and at this stage Microsoft successfully established its dominance in personal operating systems.

From the late 1980s to the early 1990s, the software industry began to boom. More than a dozen domestic operating system platforms appeared, tightly suppressing MS-DOS and taking almost the entire domestic market.

Looking back today, the reason is simple. Domestic developers knew exactly where users' pain points were, while Microsoft, far across the ocean, didn't—especially the issue of Chinese localization.

During this period, well-known domestic systems like CCDOS, Xiaojun 2.13, UCDOS, and Kingsoft SPDOS appeared. Among them, UCDOS, released in 1993, at one point held over 90% of the domestic market.

In 1991, Qiu Bojun and Li Ming developed the classic WPS 2.2 running on SPDOS, the standard in typing shops and government offices back then. (Image from the internet, contact to delete if infringing)

But to be honest, these so-called "domestic" operating systems were not developed from scratch—they were Chinese localizations of MS-DOS. In standard industry terminology, they should be called "CC-BIOS."

In other words, these systems did not take over core kernel functions like disk management, process scheduling, or file systems. What they did was intercept the underlying BIOS input/output to give DOS—which originally could only recognize English characters—Chinese character display, Chinese input methods, and Chinese printing capability.

If MS-DOS is the house itself, then CCDOS and SPDOS are just a set of painting tools for writing Chinese characters on the walls. The house wasn't built by us; no matter how good the tools, they depend on the original framework.

Of course, there were people in China who recognized the importance of the kernel. COSIX, started in 1989, chose the path of independent development from its inception.

COSIX system CD (image from the internet, contact to delete if infringing)

But this project targeted minicomputers and workstations, not personal computers, and it did not connect to the x86-compatible ecosystem that was sweeping the world.

This meant that to have a computer running COSIX, every peripheral and driver had to be custom-built. These limitations made it impossible for COSIX to break into ordinary offices and homes.

Market Competition

Many people simply attribute success or failure to "technical inferiority," but operating system competition has never been a one-on-one kernel duel. It's an all-around ecosystem war that ties together hardware, software, developers, and users.

Because an operating system is not just an operating system.

From the moment personal computers entered millions of households, up to today's "orderly" global PC world, the rules have always been firmly in the hands of Wintel!

Microsoft and Intel (image from the internet, contact to delete if infringing)

That is, Intel's x86 architecture paired with Microsoft's operating systems formed the underlying standard of the entire industry.

Every motherboard, graphics card, sound card, and printer manufacturer in the world wrote drivers with MS-DOS as the default target, and later naturally turned to Windows.

To adapt a completely new, domestically independent operating system, every peripheral would require developers to write drivers from scratch. The cost is unimaginably high, making it almost impossible.

An era-defining inflection point arrived: computers were moving from character command lines to graphical interfaces.

Windows 1.0 screenshot (image from the internet, contact to delete if infringing)

Microsoft announced Windows in 1983 and actually shipped Windows 1.0 in 1986, but it didn't achieve the expected market impact.

It wasn't until Windows 3.0 in 1990 that Microsoft's Windows gradually became the standard pre-install on personal PCs.

Windows 3.2 Chinese version Program Manager, the starting point of the graphical interface for countless Chinese people (image from the internet, contact to delete if infringing)

Then Windows 95, released in August 1995, became a milestone in the entire history of operating systems. It swept the world and took over virtually every personal PC desktop.

Windows 95 gave birth to the Start Menu and shipped with native Chinese support. (Image from the internet, contact to delete if infringing)

Meanwhile, back in China, around 1998, multiple Chinese-language Linux products began to appear.

For example, COSIX64, the first 64-bit domestic OS released by CS&S; Red Flag Linux from the Chinese Academy of Sciences; Blue Point Linux from Shenzhen XinKeEn Software Technology; and "NeoKylin" from China Standard Software and the National University of Defense Technology.

Special mention goes to Red Flag Linux. Red Flag Linux 1.0 was released in August 1999, developed by Zhongke Red Flag Software Technology Co., Ltd. At the time, it carried high hopes.

But a flower under protection ultimately couldn't withstand the wind and rain outside. At the end of 2013, the company's capital chain broke, and in August 2014 it was acquired by Wujia Wanjing.

Red Flag Linux (image from the internet, contact to delete if infringing)

At this point, the truly terrifying barrier of operating systems began to take effect: the ecosystem!

With few users, software vendors are unwilling to spend effort porting.

With scarce software, ordinary users naturally have no incentive to switch systems.

Without users, an operating system has no reason to keep iterating.

This cycle leaves domestic OS vendors utterly unable to cope.

Microsoft, on the other hand, relied on global licensing revenue, giving it a constant stream of funds to subsidize development tools and court software vendors, creating a positive feedback loop: large user base -> software adaptation -> system iteration -> continued user growth.

In contrast, China's early OS projects were mostly research tasks or fiscal projects. Once completed and accepted, they struggled to form a commercial cycle of continuous iteration and self-sustaining revenue.

The various Windows installation CDs from back then (image from the internet, contact to delete if infringing)

Not to mention Microsoft's top-tier open conspiracy: pirated Windows was almost everywhere.

This "distribution" let Windows penetrate government offices, schools, and home computers at near-zero cost, gradually squeezing out the domestic OS market.

By the time the whole society truly realized we needed a self-controllable desktop OS, Windows had already locked in the entire industry chain, making a switch impossible.

Breaking Out of the Loop

At this point you might sigh in regret: for various reasons, we missed the golden window of the PC boom. What a pity.

But now we have enough R&D capability—why not develop another desktop OS from scratch?

The answer is: the cost is too high.

Windows systems from various eras (image from the internet, contact to delete if infringing)

Building the kernel itself is actually not the hardest part.

What is truly expensive and time-consuming is the entire supporting system: graphics stacks, driver frameworks, debugging tools, development libraries, fonts, input methods, and massive amounts of productivity software across every industry.

Not to mention adaptation for all kinds of hardware: printers, fax machines, tape recorders, graphics cards, sound cards, and more.

Once a standard is established, breaking it requires exponential cost.

But you don't need to be too discouraged. If you broaden your view and step outside the single yardstick of "replacing the Windows desktop," domestic operating systems have already blossomed in many places and gained a foothold in different domains.

Servers and cloud computing were the first strongholds domestic OSes captured.

Huawei openEuler server operating system (image from the internet, contact to delete if infringing)

Huawei's open-sourced openEuler is now strong in the domestic server market, with large-scale deployments by carriers, financial institutions, and cloud vendors, and its community continues to grow.

Alibaba Cloud's Apsara, designed for ultra-large-scale clusters, is a natively distributed cloud operating system that underpins massive cloud infrastructure in China.

There are also Galaxy Kylin Server Edition, Nessus (Linx), Turbo, ZTE NewStart, and others targeting high security, industrial control, and specialized server scenarios.

Servers don't rely on fancy graphical interfaces; they value stability, security, and cluster scheduling, making them a natural fit for domestic systems.

openKylin (image from the internet, contact to delete if infringing)

The desktop hasn't been completely killed by Windows either. Under the policy of domestic substitution (Xinchuang), domestic Linux systems like Galaxy Kylin and UnionTech UOS (deepin) are gradually transitioning from "barely usable" to "convenient for daily use."

Official document processing, OA, finance, and domestic office suites are basically adapted, and they are now gradually rolling out in government and school settings.

In the mobile and IoT arena, the most significant breakthrough to date has appeared: HarmonyOS.

It has a self-developed microkernel and follows a distributed architecture, connecting phones, tablets, car systems, and smart home devices. It is now the third-largest mobile operating system in the world.

HarmonyOS's goal is clear: through distributed methods, along the new track of cross-device collaboration, it aims to build an ecosystem independent of Wintel and Android. Initial results are visible.

HarmonyOS (screenshot from HarmonyOS official website)

In addition, there is a large group of embedded real-time operating systems that ordinary people rarely hear of but developers know well: RT-Thread and SylixOS. They run widely in industrial equipment, robots, and home appliances.

Conclusion

We need to face reality objectively: the desktop ecosystem gap remains the biggest challenge. Porting professional productivity software, niche industry tools, and games will be a long and hard battle.

Looking back at this history, we shouldn't fault the researchers of that era. CCDOS, 2.13, and SPDOS are not failures; they excellently fulfilled their historical mission, bringing Chinese characters to PCs, giving rise to WPS, and supporting the foundation of domestic informatization in the 1980s and 1990s.

They were solutions for the era, just not the next-generation general-purpose operating systems.

Today's road for domestic operating systems is no longer about "replicating a Windows."

One path is to continuously strengthen Linux-based distributions, holding steady in government, state-owned enterprises, and industrial scenarios.

The other path relies on HarmonyOS to seize the new track of the Internet of Everything, striving to build our own ecosystem before new standards take shape.

I'm Li Jianyi, a programmer who looks back at the turbulent history. Welcome to follow my WeChat public account of the same name, "Li Jianyi."