China DOS Union

Take a look at the current China DOS Union, which has become so desolate. All day long, it's just discussing some outdated topics and obsolete technologies. Especially in this development section, old posts are frequently brought back to the top, but new posts rarely appear.

At the same time, the development of FreeDOS is also extremely slow, lacking vitality and energy.

I don't know when I read a sentence: A system lacking developer support is a system without a future (paraphrasing).

Anyone with a little experience in using DOS knows that DOS has many limitations and is very inconvenient to use. The fundamental reason is its 16-bit kernel. Although 32-bit programs can be used through DPMI and DOS extenders now, various drivers, Chinese systems, etc., that we often use are still 16-bit TSR programs (these programs must be 16-bit TSR). Although more memory can be used for data exchange through XMS, the active space of the program is still limited within a 1M space. This undoubtedly seriously hinders the development of DOS.

This problem is a major weakness of DOS. I think as long as this problem is not solved, DOS cannot have better development. In fact, currently, except for some bottom-level work very close to hardware and in embedded fields, DOS is basically useless elsewhere, especially in the desktop field.

Let's first take a look at Win9x. It is an operating system running on DOS. Of course, some people think it's just a program on DOS. I won't discuss that. In any case, we can see that Win9x can achieve true multitasking. Even its drivers are all 32-bit. Of course, it can also make a program run in the background (and such background running is completely different from TSR, it's a 32-bit multitasking mechanism, not limited by 1M memory). This makes me think that we should be able to develop a kernel on DOS, write programs on this kernel, so that we can completely break through the 16-bit limit, develop pure 32-bit mouse drivers, IFS drivers, Chinese systems, etc., completely abandon TSR, and no longer worry about the memory limit problem.

So there are two paths: 1. Be compatible with the Win32 API, just like what the current HX Extender does. In this way, development will relatively get more user support, and there are a large number of software supports. But since the Win32 API is closed and very complex, the development difficulty is relatively large, and it's difficult to be completely compatible. 2. Develop completely independently. In this way, development will be relatively easier, but there are no ready-made programs running on it, and all software must be redeveloped. At least they need to be recompiled. Of course, we can still develop software for being compatible with the win32 API on this new kernel (just like wine on Linux).

However, we also have a third path, which is to develop a pure 32-bit kernel DOS system. If it can achieve full direct access to hardware and can well be compatible with various existing DOS programs, it can be regarded as a DOS system. For now, if we want to take this path, participating in FreeDOS-32 is the best choice (FreeDOS-32 has been stagnant for more than 3 years).

Among these three paths, no matter which one is chosen, very high-level bottom-level development is required. Even the first two paths are actually no different from developing an operating system kernel. And I can't say which path is the best and easiest to succeed, but at least one of these three paths is the correct path, because only by moving towards 32-bit will there be a future.

In addition, for the internationalization of DOS, the participation of Chinese people is also needed. Chinese is also one of the important international languages! And if the 16-bit limit is broken, we can introduce Unicode to DOS, making it support not only Chinese but also languages all over the world. And the display of text requires GUI technology. We can develop something similar to the X Server on Linux to support graphics, but in a 16-bit environment, developing such a program is almost impossible.

The above are my humble opinions. Everyone is welcome to discuss! But if this post gradually sinks, I will feel sad, because that will indicate the real decline of DOS - no one is willing to develop for DOS anymore.

[ Last edited by brglng on 2008-10-2 at 20:14 ]

Instead of laboriously developing 32-bit, it's better to directly use Linux. The latter is much better in terms of maturity, driver support, etc., and there are also many source codes to draw on.

Originally posted by netwinxp at 2008-10-7 04:03 PM:
Instead of working hard to develop 32-bit, it's better to directly use Linux. The latter is much better in terms of maturity, driver support, etc., and there are also many source codes to draw on.

I have also thought about this problem. Linux is certainly very good, and I myself use Linux as my desktop system. But I think that DOS still exists so far for its irreplaceable places. For example, it has direct and unrestricted access to hardware, which is very useful in many aspects. And when studying microcomputer principles and assembly language, DOS is often indispensable. It is also very useful for us to study computers. And who knows that after transforming into 32-bit, more new uses cannot be discovered? Everything can only be spoken by facts. We haven't developed it, haven't used it, how can we know whether it is good or not?

In fact, I think that the current decline of DOS is more because it itself is not advanced enough, but because of Microsoft's monopoly, closure and its abandonment of support for DOS. In fact, Win9x has greatly expanded the original DOS functions. If it opens source code, it must have great development. And Microsoft does not support DOS, finally turns to NT, and does not develop 32-bit DOS. Because of its monopoly position, DOS is thus abandoned.

As far as I know, Unix was also an early 16-bit operating system. It was only after the appearance of 32-bit processors that there was 32-bit Unix. And at this time, Microsoft did not develop 32-bit DOS, but turned to OS/2 and NT, and the source code of DOS is not open, so DOS declined.

[ Last edited by brglng on 2008-10-9 at 15:48 ]

I want to ask a question, what functions of DOS are we using now?
Graphics? Basically rely on direct screen writing.
Memory? Without DOS, more can be accessed, and you can let whoever stay resident, no trouble at all.
Disk? Now the access to FAT/FAT32 basically doesn't need to rely on DOS.
To be honest, if you want freedom, it's more convenient to directly use the boot program to call your BIN program, and the switching between protected mode and real mode becomes simpler, and it's less likely to conflict and have key data segment reentry. DOS actually mainly has INT 21H, and it's almost useless now.
Linux you can let your program run in KRNL, which is equivalent to an unrestricted 32-bit or even 64-bit system, and there are many source codes to "copy", which is more rewarding than Windows' "building with blocks".
PS: As memory breaks through 4G and hard drives are about to reach more than 2T (the traditional partition table architecture is facing being replaced by GPT), the necessity of using 64-bit systems is becoming more and more urgent. Even if 32-bit DOS is developed, it will soon face the situation of being eliminated.

[ Last edited by netwinxp on 2008-10-9 at 22:09 ]

Originally posted by brglng It's basically useless elsewhere, especially in the desktop field.

Just right, the desktop field needs to use DOS. Use this environment to upgrade and refresh the motherboard BIOS, including server motherboards. Now there are also those used to upgrade BIOS under Windows, but there are more people upgrading motherboard BIOS under DOS. Those who don't know how to use DOS well are also upgrading motherboard BIOS under DOS. If the motherboard in the future can't be upgraded BIOS under DOS, then DOS will really be eliminated.

Regarding 64-bit, I have also considered it. It is very likely that in the near future, processors will no longer support 16-bit, and then the current DOS will truly perish.

However, don't look at DOS being not focused on in China. There are still quite a number of people abroad (of course, not as many as Linux) developing for DOS. Also, there is software like DOSBOX on Linux for emulation. This shows that DOS still has many uses.

I always feel that DOS has not completely disappeared so far, and there must be something irreplaceable about it. Just like what the person above said about upgrading BIOS. This is just an example.

Moreover, DOS was also a commonly used desktop system, which shows that using DOS as a desktop system is not impossible.

For 64-bit, I think as long as 32-bit is developed, it is not impossible to develop 64-bit.

Finally, I still want to emphasize, how can you know whether 32-bit DOS is good or not if you haven't used it? It's just like someone who has only used Windows and has never used Linux won't know where Linux is more advanced than Windows.

>And the display of text requires GUI-related technologies. We can develop something similar to the X Server on Linux to support graphics, and in a 16-bit environment, developing such a program is almost impossible.

It's not impossible, heh. We've already made it.

It's a pity I don't know programming, can only look forward to,...

ms-dos 6.22 doesn't have source code, does it? Microsoft even gave the source code of Windows to the Beijing government, and that hacker got it out.

[ Last edited by xiaoyunwang on 2008-10-11 at 15:41 ]

Originally posted by brglng at 2008-10-11 12:32:
Regarding 64-bit, I also have considered it. It is very possible that in the near future, processors will no longer support 16-bit, and then the current DOS will truly perish.
But, don't look at DOS not being paid attention to in China, there are still relatively ... in foreign countries

Currently, desktop CPUs are AMD64 (X86-64 is compatible with its standard), which itself is compatible with 64/32/16 bits. Let's not mention IA64 which is incompatible with everything.
Does 32-bit DOS need to use protected mode? Then in protected mode, to run DOS programs, 8086 virtual mode should be used. Unfortunately, it has great restrictions on privileged instructions such as I/O. Then the convenience of directly accessing hardware is gone. Also, the original BIOS interrupt services in real mode cannot be directly used now... and so on will make all the original conveniences gone. Then should it still be called DOS32?
Actually, most of LZ's ideas have already been in WIN9X. Of course, it also has slightly restrictions on hardware access. A powerful and relatively stable system will necessarily lead to more restrictions. For example, we use I/O-level programs which are basically dominated by the program with almost no restrictions; programs using BIOS interrupts need to abide by the regulations of BIOS interrupt calls; even the memory arrangement when using DOS has to be handled by DOS. Don't make it so that conventional memory is insufficient and cannot run. When compiling an interrupt, one has to prevent reentrancy problems...

[ Last edited by netwinxp on 2008-10-13 at 00:06 ]

Bump

Bump

Instead of struggling to develop 32-bit, it's better to directly use Linux. The latter is much better in terms of maturity, driver support, etc., and there are also many source codes to draw on.

There are indeed very few posts in the forum, hey