DigestII've been playing with DOS for about 20 years. I should have known about DOS from the end of 1985 or the beginning of 1986. At that time, DOS was the best operating system for personal computers.
Now I'm still doing things under DOS, mainly on 86-architecture machines, based on DJGPP + ALLEGRO. I also use TURBO C for some simple tasks. I rarely use assembly language now, only occasionally when I have to.
The advantage of DOS is that it can operate hardware without any scruples and is extremely fast. So my programs basically run on the motherboards I made specifically. For example, the newly developed motherboard for a project uses an AMD gx3 chipset, the bridge uses CS5536, is equipped with an ALC202 AC97 sound card, an bt878a image capture chip, and connects an AIT2138 image decoding chip and a DS3231 high-accuracy clock chip via I2C on CS5536. When DOS is started, it's all directly programming the chips, which is cool to run!
The trouble with DOS is its non-reentrancy, which makes it necessary to be very careful when building multitasking, not being able to do as one pleases. But there's no need to modify DOS for this because DOS 6.22 nowadays is really very stable. If you change it, who knows what might happen. When I was young, I once completely changed DOS in a project, from the keyboard, display to disk operations. As a result, none of the debugging tools could support it, and that project really wore me out.
I don't know how many people are really developing under DOS now. A foreign project a couple of years ago first used WINCE. The foreigner said it was unstable and not reliable after using it. Then I redid it with DOS, and the foreigner kept saying it was good. Originally, it's just to start when you want to and shut down when you want to. Using a 200 CPU, the real-time performance of image capture is extremely good. What else is there to say.
Now I'm still doing things under DOS, mainly on 86-architecture machines, based on DJGPP + ALLEGRO. I also use TURBO C for some simple tasks. I rarely use assembly language now, only occasionally when I have to.
The advantage of DOS is that it can operate hardware without any scruples and is extremely fast. So my programs basically run on the motherboards I made specifically. For example, the newly developed motherboard for a project uses an AMD gx3 chipset, the bridge uses CS5536, is equipped with an ALC202 AC97 sound card, an bt878a image capture chip, and connects an AIT2138 image decoding chip and a DS3231 high-accuracy clock chip via I2C on CS5536. When DOS is started, it's all directly programming the chips, which is cool to run!
The trouble with DOS is its non-reentrancy, which makes it necessary to be very careful when building multitasking, not being able to do as one pleases. But there's no need to modify DOS for this because DOS 6.22 nowadays is really very stable. If you change it, who knows what might happen. When I was young, I once completely changed DOS in a project, from the keyboard, display to disk operations. As a result, none of the debugging tools could support it, and that project really wore me out.
I don't know how many people are really developing under DOS now. A foreign project a couple of years ago first used WINCE. The foreigner said it was unstable and not reliable after using it. Then I redid it with DOS, and the foreigner kept saying it was good. Originally, it's just to start when you want to and shut down when you want to. Using a 200 CPU, the real-time performance of image capture is extremely good. What else is there to say.
www.ecgui.com