中国DOS联盟

I personally think that the GUI under DOS is of no significance! The strength of DOS does not lie in graphics. Moreover, since one chooses to use DOS, it is definitely not because of its good GUI. For embedded systems, functionality is often the primary concern, interface comes next, and within the interface, functionality is primary, and aesthetics come next. In the projects I have done, I can hardly remember any that have a mouse, and most don't even have a keyboard. I have a general setting solution in my projects, which is to use a three-button small button, sometimes on the panel, sometimes externally connected, and plug it in when needed. All settings are done with three buttons. Think about it, what is the use of a GUI in such a system. When we do LINUX embedded systems, we need to cut down a lot on LINUX, and basically no one will keep the GUI.

First of all, I want to say that I really put in a lot of effort and it's very good to see the GUIs under DOS made by some young people. I just, based on my 20 years of experience using DOS, think that it may have a bleak future and no practical value. Just my personal opinion.

Look at the development history of WINDOWS. In 1985, MS launched WINDOWS 1.0, and in 1987, 2.0. These two versions basically had no vitality. Maybe many people don't know. In 1990, WINDOWS 3.0 appeared, and WINDOWS began to rise from then on. The 3.1 version of WINDOWS in 1992 was a very representative version that made WINDOWS popular. But until before win2000, WINDOWS was just an application under DOS and didn't count as an operating system at all. To put it bluntly, it was just a GUI under DOS. I want to say that developing a GUI under DOS is just following the old path of MS, doing what they have done, and it's really unnecessary.

In any case, it's恐怕 very difficult for anyone to catch up with making the interface look like Windows and MAC OS. Moreover, I think that if there are very high requirements for the GUI, then generally the requirements for stability and reliability will not be high. So it's completely okay to not use DOS and use WINDOWS or WINDOWS CE, which is better, and it's easier to develop.

Once again, just my personal opinion.

If it is for DOS, what the original poster said is very correct; however, if the software structure is reasonably done, the separation of the underlying drivers is thorough, and the core abstraction layer is comprehensive, the portability should be relatively good, and it can be easily transferred to other platforms...

Everything is done very well, and portability is also very good. A GUI was made under DOS and successfully ported to other platforms. What is the significance? We can't do something just for portability, and we can't do a GUI just for the sake of a GUI. The thing made must have a demand. After a thing is completed, there must be thousands of people using it in various different scenarios. Only then can you proudly say that it is basically without problems. Testing your own thing alone will never make a good thing. So, demand is the driving force for the success or failure of a thing. As for DOS, there are still people like me using it in practice nowadays, and many people like my customers tell me that what you did under DOS is better than what they did under WINDOWS. This is the necessity for DOS to continue to exist. But for so many years, DOS has almost stood still. Why? Because there is no corresponding demand. What should DOS develop into? After thinking it over, it's nothing more than WINDOWS and LINUX, which already exist. Why do we still need to do it? Why can DJGPP develop? Because with the improvement of machine configuration, people need to run 32-bit code. Why can ALLEGRO develop? Because people hope to display pictures as well as under WINDOWS on low - configuration machines, and at this time, it's very difficult to do it in pure DOS. All these are driven by demand!

Why did Linus develop LINUX? Because some performance of MINIX was not satisfactory, and there was no alternative operating system, which is also demand.

No thing can be successful without demand. When you want to do something, think about how many people have this demand, and ask those people related to the thing you are doing whether they need the thing you are making. Maybe these will help you make the right decision.

In the past, I used an operating system for real - time on the X86 architecture, an INTEL one called iRMX. It had very good performance. We used it to make a ten - thousand - door program - controlled switch, and it passed the monitoring of the Ministry of Posts and Telecommunications and got the access license. I always miss that operating system. Unfortunately, many years ago, INTEL had announced that it would no longer support this system. I think it's not because the system was bad, but because the number of users was too small, which was not enough to make it have a vigorous vitality.

（1）If this "embedded system" is an X86 architecture PC104 computer, with a CPU of 8086, 80186, 80286, 80386, 80486, and using LCD or VGA
（2）If you are currently an electronic development engineer or an embedded project development engineer
（3）If you are the boss of an enterprise

----If you choose to use "WinCE" or "Unix", then in the domestic market in China, with the current market conditions (2007, Shenzhen), you need to pay at least ￥8000 RMB to hire software engineers.

----If the CPU is 8086 or 80186, you can only choose DOS, and it is the only option.

----If the CPU is above 80386, if you do not choose DOS, you can only choose Unix or WinCE. Then, in China, are there more people proficient in DOS development or more people proficient in Unix or WinCE?

In embedded applications, choosing the "DOS" system is the most "cost-saving" solution, including development time control, cost control, etc.

Currently, embedded PC104 generally uses an electronic disk of 16M (or more) or a CF card of 16M (or more). In my embedded application, I have installed the development tool BC31, so as to modify the program on site.

Nothing is all-purpose. Precisely because the hardware of embedded systems varies widely, the need for a GUI under DOS has developed. It is not meaningful to always discuss DOS side by side with WINDOWS and UNIX.

In short----Different people have different preferences; DOS, UNIX, and WinCE each have their own advantages. And in embedded applications, it is often based on the technical accumulation of the enterprise itself, so as to make a wise choice of which operating system to use.

I have seen many "people proficient in UNIX, and not not proficient in DOS programming".
I have seen many "people proficient in DOS, but with only limited knowledge of UNIX from written materials".
I have seen many "people proficient in DOS, but have never heard of WinCE".

Last edited by firstsail on 2007-6-3 at 07:44 PM ]

The key is that now people will buy it when they see your interface is very good, and they will only understand the actual functions after using it. Salespeople will all say that the functions are no problem. So the first thing to present to others is the interface.

If you choose to use "WinCE" or "Unix", then in the domestic market of China (Shenzhen in 2007), you need to offer at least ￥8000 RMB to hire a software engineer.

Friend on the 4th floor, may I ask, if hiring a developer under DOS, what is the monthly salary?

Last edited by enjoyer on 2007-6-5 at 11:34 PM ]

Hiring DOS developers, the monthly salary is only 2/5 of that for UNIX developers, approximately ￥3500 (2007, Shenzhen). If one is proficient in both electronics development and DOS development, then ￥10000 is no problem (2007, Shenzhen)

Last edited by firstsail on 2007-6-4 at 01:16 PM ]

I think the friend on the 4th floor misunderstood my meaning. I read my post again. I didn't mean to compare DOS with LINUX and WIN CE. I just feel that so many people are pondering the GUI under DOS, and they seem to be going the way MS once did. I have been developing products under DOS. Personally, I am more familiar with the 8086 architecture. Only in recent years have I worked more in the protected mode of 386. I also need to make interfaces under DOS. But from my development experience, most products face a very small LCD. In real mode, the graphics functions of BC or TC are enough to make a satisfactory interface. There is no need for multi-window at all, and there is no need for support such as keyboard and mouse. The actual resolution used is often only 320 * 240, and sometimes a low-resolution character mode is also used. Even if developing under VGA, VESA is enough to make a satisfactory interface. Why go to the trouble to develop a GUI under DOS that looks a bit like WINDOWS? I can't think of where I would use it. If developing in protected mode, the image and interface capabilities provided by ALLEGRO far exceed my needs. I think only those who make games under DOS would need such complex graphics capabilities.
If only an occasional project needs a multi-window GUI similar to WINDOWS, making it into a product probably won't have much vitality.
In fact, if someone develops library functions for displaying JPEG or GIF under DOS, or converting BMP to JPG, or JPG to BMP, that would be useful. Once I needed to store captured images in JPG format in a project, and of course, I needed to display JPG format. I searched for a long time and didn't find a library, so I had to make it myself, which was very tiring.

In addition, I also want to say that for making windows, it's not a big deal if you don't understand hardware. For making DOS, if you don't understand hardware, I really don't know what he can do. Sometimes, software engineers must propose good design methods for hardware engineers. Take the latest motherboard of GX3 + CS5536 I opened as an example. I asked the engineer to connect 2 devices on the I2C of CS5536, one AIT2138 (VGA to AV) and one DS3231 (high-precision clock). The operations on these two chips via I2C were all normal. But if there is the AIT2138 chip, the WATCHDOG of CS5536 will be a bit abnormal. I checked the circuit, and everything was in accordance with the specifications. To be honest, I still don't know why this problem occurs. But I asked the engineer to connect AIT2138 to the I2C of BT878A or to the idle GPIO, and there was no problem. This kind of thing, software engineers can't solve, and hardware engineers didn't make a mistake. If there is no one who is proficient in both software and hardware, disputes are inevitable.

From my development experience, the vast majority of products are faced with a very small LCD. In real mode, the graphics functions of BC or TC are sufficient to make a satisfactory interface. There is no need for multiple windows at all, and there is no need for support such as the keyboard and mouse

The above situation indeed means that "Windows-like GUI" is not very useful. Such projects can often be replaced by "microcontrollers".


Each embedded project has different hardware architectures, and market segmentation needs to be done. Regarding the development of GUI under DOS, I think that only the discussion of "embedded applications" under the X86 architecture has practical significance!

Nowadays, people have been deeply rooted in the Windows interface, and even factory "operators" can use it well. If you choose "DOS" as the embedded operating system and need a good interface, the "Windows-like" GUI has great advantages. Otherwise, I feel that the product interface always looks as complicated as the dashboard on an airplane.

Smart instruments mostly use CPUs below 80486 and are equipped with TFT screens of 320*240 or above. At this time, choosing Dos + Windows-like is a very good choice. College graduates before 2001 basically were familiar with DOS, and there are a large number of talent reserves in China.

Under the same hardware architecture, the instrument using "Windows-like GUI" must be able to sell better and have stronger competitiveness than the instrument not using "Windows-like GUI", thereby bringing more product added value to the enterprise! Why not?

Last edited by firstsail on 2007-6-4 at 06:40 PM ]

Haha, don't argue, everyone has their own reasons.
I've done a lot of work under DOS, and also a little work under Windows, so I think everyone is reasonable.

Asking the friends upstairs, what kind of projects do people who are proficient in DOS development and also understand electronics development mainly do? I don't understand the application of DOS in the embedded field. Please give me some pointers, thank you!

（1）Most CNC machine tools
（2）Most medical equipment
（3）Digital oscilloscopes
（4）Smart instrumentation
（5）Testing equipment
（6）PLC (many PLCs use 8086 CPU internally)
（7）Lower - level machine parts in large systems
......




.......

The spirit of your argument is worthy of my learning, practical and seeking truth.
But both of you are correct. One focuses on simplicity and has practical application value, and the other is dedicated to research, study, and has his own pursuit, which can reflect the spirit of gun. If only I could have the technologies of both of you. Then I could earn money to support myself.
Your words have taught me a lot. I hope you can also show me a way, for me, a graduate, not to live on a tight budget every day.

Do you have a complete set of programs? Including mouse, keyboard, graphical interface,
I'm in the learning stage, thank you very much.
Also, I will graduate this year. Are there any employers in Shenzhen hiring fresh graduates?

Alas, my admiration for the original poster is like a surging river. Recently I want to do an embedded project under DOS, but I find there are very few materials online and don't know how to start.