16
Wang Fei went off to play cards again, just at that small gathering spot next door—it really wasn’t far. Today was Saturday, and Wang Fei and her group of girlfriends had a rare chance to get together. They were drinking and having fun, with music blasting at full volume.
That room was quite large, enough for them to do whatever they wanted. A lot of thought went into soundproofing during renovation, so Wang Xu wasn't too bothered on his side.
Wang Xu himself was leisurely writing software proposals, ranging from mathematical software to 3D engines, plans for all sorts of software.
Wang Xu truly had no solutions for hardware; that stuff required too much and too complex equipment, so he could only take his time with it.
But in terms of software, he could still overtake others, and he didn't even need to play any curveballs; he could just overtake them in a straight line.
The problem with software, especially the first version, is usually not about development, but about commercialization.
DOS was not the best operating system, but it was the most successful operating system in terms of commercial operation.
That's why it was continuously improved and eventually defeated countless competitors to become the dominant operating system.
Wang Xu didn't plan to go the desktop system route; no one could beat the Wintel alliance in that path. In this era, it was an unbreakable alliance.
Therefore, he could only find another way, taking the server and workstation route.
At least in the 1990s, servers and workstations were still the domain of Unix.
Unix is not a unified operating system, but a general term for a series of operating systems.
The Unix versions from various companies were not completely compatible, and porting software between them was quite troublesome. Wang Xu was taking the Linux route.
Richard Stallman founded the GNU project at the MIT Artificial Intelligence Lab (AI Lab), aiming to create a completely free Unix-like operating system.
In this project, Richard and other volunteers developed the GNU toolchain (GCC, GDB, GNU Make, etc.), but the operating system kernel was never completed.
Linux, on the other hand, was a self-entertainment project created a year ago by Linus Torvalds, a genius student from the University of Helsinki, Finland.
It was actually an exercise based on the GNU toolchain, initially only dealing with the basic kernel and I/O parts, which was quite incomplete.
However, Linus, as a genius, had very advanced ideas. The microkernel, replaceable components, and the new operating system concept where everything else was an external application, made this operating system stand out from many other exercises.
Later, he open-sourced this kernel on the internet, making it available to the whole world. He also explained his development philosophy in the accompanying files.
He also allowed anyone to modify and distribute it, and even broke out of the old framework where free software did not allow commercial operation, allowing this operating system to be commercially operated for the first time.
Free, commercially usable, and with advanced concepts, this operating system quickly gained the recognition of countless programmers, who voluntarily participated in its modification and updates.
With programmers from all over the world joining, this operating system finally had its first stable kernel.
Probably this year, or maybe it already has, the first distribution version of Linux will be packaged by a programmer named Patrick Volkerding, and only then will Linux truly be on track.
However, China hasn't connected to the Internet yet, so Wang Xu couldn't obtain the Linux source code at all. Moreover, Wang Xu genuinely didn't think much of the current Linux version.
Aside from its advanced concepts, the code itself didn't have anything particularly outstanding.
Wang Xu's operating system plan actually started earlier than Linux. It was the earliest software product developed within the Supercomputing, so early that young teachers had already started development by the end of last year.
Currently, it is only adapted for Intel series chips and can only support a few types of graphics cards and display modes, with very limited support for various peripherals.
Of course, this was already enough, as most computers within Tsinghua University were purchased in bulk, so only a few types of graphics cards were needed.
The system was also unable to call multiple external GPUs to achieve Wang Xu's original "Supercomputer" dream.
However, a rudimentary, single-PC-only, command-line-only, application-less, and even buggy first development version still appeared in March.
Wang Xu was not very impressed with this thing, so he didn't make a fuss or publicize it widely. Although Professor Yan really wanted to boast about it, Wang Xu persuaded him to wait a bit. Anyway, the complete version was not far off, so they could boast about it together once it was ready.
But, whether publicized or not, China's first truly meaningful desktop operating system appeared so quietly.
Wang Xu simply named it Cos, an abbreviation for Chinese OS, without using any mythical beast names.
Yes, that's right, Chinese Operating System. Although the first version didn't support Chinese, its 32-bit underlying logic allowed it to easily extend support for various languages.
Of course, you can also understand it as an operating system for Chinese people.
This operating system currently only has some basic functions: it can boot up, and some small tools have been ported for development. It doesn't even have basic network connection functions, the reason being that the corresponding network programs haven't been developed yet.
Wang Xu directly skipped the 8-bit and 16-bit eras; the kernel was directly 32-bit. The result was that the final debugging and testing of this product could only be completed on machines above 80386.
To solve this problem, Wang Xu racked his brains again and created a compatibility mode, allowing 16-bit computers to also run this operating system.
However, the speed was truly nothing to write home about; aside from experiencing it, it had no practical use.
Nevertheless, development could still be done on older machines, as GNU itself could adapt to 8-bit, 16-bit, and 32-bit compilation modes; one just needed to change a parameter during compilation.
This project was very difficult from the very beginning, both for teachers and students, as very few people in China had actually developed an operating system.
Even the teacher who wrote the operating system textbook only knew the theory and had never actually implemented an operating system. The complexity and trouble involved were imaginable.
It was only because Wang Xu's project plan was detailed enough, and he even personally solved many difficult problems, that the first version of Cos was successfully born.
What's more, in the beginning, all development could only be done on the DOS operating system; there wasn't even Unix.
Then, using rudimentary editing tools, developing something that couldn't even run on DOS—you can imagine how tedious that was.
However, with the official completion of the first version of Cos, the development work finally got on track. Finally, the teachers and students who had suffered for so long could develop and debug on one machine.
This was all thanks to the porting of some common Unix tools, such as cat and grep.
The source code for these things could actually be found; senior students who had studied abroad had brought them back long ago, so Wang Xu didn't even need to get involved.
—
(See you tomorrow)
