中国DOS联盟论坛

China DOS Union

-- Unite DOS · Advance DOS · Grow DOS --
Union site: www.cn-dos.net Forum site: www.cn-dos.net/forum
Guest | Log in | Register | Members | Search | China DOS Union
中国DOS联盟论坛
The time now is 2026-10-31 19:15
47,815 topics / 349,930 posts / today 0 new / 48,290 members
DOS开发编程 & 发展交流 (开发室) » Recently, some preliminary research and development achievements in DOS development
Printable Version  4,518 / 33
Floor16 sunx Posted 2007-07-28 01:19
初级用户 Posts 17 Credits 38
It is estimated that it can still run without removing com, lpt, pci, DOS, as long as the basic interface of the BIOS does not change. Because in the oldest machines, com and lpt were not originally there, and they had to rely on multifunctional cards (ISA expansion cards) to be added. And there were originally no pci.

Basic is indeed a good thing. For the recent DOS projects, if it weren't for more operations on data structures, I would have chosen something like PowerBasic.

As you said, the biggest attraction is that it is fast to write.

If the hardware meets the requirements to run the Windows series, it is better to use Windows. If you can use Windows, I would never go back to doing DOS. It's still because it's fast to write, but the fact that Windows is paid is a pity.

Originally posted by netwinxp at 2007-7-27 23:30:
As for industrial control, in the early 1990s, there were many based on MSDOS, DRDOS. In recent years, those based on Windows have accounted for the majority, and more and more are using VB + AD/DA acquisition card driver programs (fast development speed), this...


[ Last edited by sunx on 2007-7-28 at 01:31 AM ]
Floor17 windowsvesta Posted 2007-07-28 10:24
初级用户 Posts 67 Credits 138 From 云南
Originally posted by netwinxp at 2007-7-27 11:30 PM:
As for industrial control, in the early 1990s, there were many based on MSDOS and DRDOS. In recent years, those based on Windows have accounted for the majority. More and more applications use VB + AD/DA card driver programs (because of fast development speed)...


For many automation projects, the host computer is generally "Windows", and the slave computer is generally UNIX or DOS.

When directly using an AD card on Windows, the AD card must have FIFO function, otherwise the sampled data will lose time segments, because Windows is not a real-time operating system, and it is normal to have delays of 20ms to 500ms from time to time. High-precision instruments and meters are generally: single-chip microcomputers, ARM, IPC/PC104 using DOS, IPC/PC104 using UNIX, IPC/PC104 using WinCE. Only ordinary automation applications are Windows, that is, the so-called "system integration" in general, and it is difficult to break through 1% accuracy.

[ Last edited by windowsvesta on 2007-7-28 at 11:47 AM ]
Floor18 netwinxp Posted 2007-07-28 11:00
高级用户 Posts 366 Credits 741
For lower - level machines with relatively strong real - time requirements, smart chips are probably more appropriate. After all, the clock of a PC will drift and is not so accurate. And smart chips usually come with their own development systems, and DOS probably can't help much either.
Floor19 windowsvesta Posted 2007-07-28 11:53
初级用户 Posts 67 Credits 138 From 云南
I have seen a system developed by Americans, with a selling price of $1 million, equivalent to about 7.6 million yuan today. It adopts a dual-host structure, with the upper host being Windows NT and the lower host being DOS. Two industrial control computers and one monitor are used. The architecture is

Hardware ---> DOS ---> Window NT
| |
|---------------------------

[ Last edited by windowsvesta on 2007-7-28 at 12:06 PM ]
Floor20 windowsvesta Posted 2007-07-28 12:28
初级用户 Posts 67 Credits 138 From 云南
For DOS programmers,

The number of Basic programmers > The number of C language programmers > The number of C++ programmers.

So,
------Developing projects using Basic language is for entry-level.
------Developing projects using TC + C language is for mass-level.
------Developing projects using BC31 + C++ language is for R&D-level.
------Developing projects using DJGPP + C++ language is for elite-level.


So the owner should not forget the "pyramid"-shaped distribution.

[ Last edited by windowsvesta on 2007-7-28 at 12:55 PM ]
Floor21 windowsvesta Posted 2007-07-28 12:41
初级用户 Posts 67 Credits 138 From 云南
Turboc C encapsulates TCC
Borland C++ encapsulates BCC

I don't know when the new version of DJGPP can encapsulate GCC. If it encapsulates GCC, then replacing BC will be just around the corner. When that day really comes, I will unhesitatingly abandon BC and run to DJGPP, but now it's "still too early".

[ Last edited by windowsvesta on 2007-7-28 at 12:54 PM ]
Floor22 shanliang8008 Posted 2007-07-28 23:41
银牌会员 Posts 514 Credits 1,148 From 河北
Floor23 windowsvesta Posted 2007-07-29 15:36
初级用户 Posts 67 Credits 138 From 云南
Saying that GCC, a tool with such a strong influence now, is something that only people under Linux would be forced to use



Suppose there are 100 IT people. Among these people, 80 have never heard of GCC, 10 have only heard of GCC, 4 have used GCC, and only 1 uses GCC regularly. So the ratio is 99:1.

[ Last edited by windowsvesta on 2007-7-29 at 03:38 PM ]
Floor24 sunx Posted 2007-07-30 03:48
初级用户 Posts 17 Credits 38
to windowsvesta:

Arguments cannot be made by whim, and even less can it be "because it's okay so it's okay"

The following are several data for your reference

The most mainstream VB only had a 21% usage rate in 2001
Apache has a 52% market share in web servers, and even reached 70% in 2005

Hope you understand what these two data represent


I don't object to debate
But you are in technology, you should be more rigorous

[ Last edited by sunx on 2007-7-30 at 04:19 AM ]
Floor25 netwinxp Posted 2007-07-30 06:55
高级用户 Posts 366 Credits 741
I agree with the person above. We technical people can use theories as evidence, can also use actual data as evidence, and can also use counterexamples to refute, but we can't use assumptions as evidence at all.
Floor26 windowsvesta Posted 2007-07-30 09:28
初级用户 Posts 67 Credits 138 From 云南
When MSN first logged into China, everyone thought it was just another relationship like "Office" and "WPS", and it could completely sweep "QQ" into the trash. The fact is, if you go to an internet café, the instant messaging tool is still QQ, and there are basically zero people using MSN in internet cafés. The user group of MSN in China is just "fake foreigners" and "white-collar workers".

Because QQ already has a huge user group, even if the market share "calculated by percentage" decreases gradually every year, the absolute number is still increasing. Does QQ have such high technical content? Is there no second IT enterprise in the whole country that can do it? But why no one is willing to compete with QQ? Because that is destined to end up "in a bad way".

A good technology doesn't necessarily mean it's the best in the market.

The impact of DJGPP on BC is just like the impact of Linux on Windows. No matter how loud the slogan is, ten years from now, BC will still be the overlord under DOS. Count with your fingers, Linux has been around for sixteen years, and whether it has knocked Windows down or not is another matter...

BC and DJGPP are not opponents of each other, but complement each other. DJGPP is a supplement to BC, an advanced application of BC. That is, when BC can't meet the needs, then DJGPP comes into play. The vivid metaphor is "BC is the first echelon, DJGPP is the reserve echelon". That is

if (BC meets the needs)
{
}
else if (DJGPP meets the needs)
{
}
else // others
{
}




This is my view!

[ Last edited by windowsvesta on 2007-7-30 at 12:00 PM ]
Floor27 windowsvesta Posted 2007-08-02 12:39
初级用户 Posts 67 Credits 138 From 云南
Originally posted by netwinxp at 2007-7-28 11:00 AM:
For lower-level machines with strong real-time requirements, intelligent chips are probably more appropriate. After all, the clock of a PC may drift and is not that accurate, while an intelligent chip usually comes with its own development system, and DOS seems to be of little help.



Generally speaking, "intelligent chips" are used for instrumentation where the "quantity market" or the volume and cost of the X86 architecture are too high. In complex systems, it is often used as a lower-level machine of the lower-level machine.

Generally, the following two-level architecture is adopted

Intelligent chip instrumentation-----|
Single-chip microcomputer type instrumentation-----|
Sensor--------------------|-----------------X86 architecture + Windows
...... ----------------------|
Other industrial field signals-----|



When complexity is relatively high, the following three-level architecture is generally adopted

Intelligent chip instrumentation-----|
Single-chip microcomputer type instrumentation-----|
Sensor--------------------|-----------------(Multiple) X86 architecture + DOS/UNIX/WinCE---------------X86+Windows
...... ----------------------|
Other industrial field signals-----|


For intelligent systems involving "a large number of real-time drawing needs", the X86 architecture + DOS/UNIX/WinCE can better comprehensively meet the requirements.

[ Last edited by windowsvesta on 2007-8-2 at 12:50 PM ]
Floor28 john12900 Posted 2007-08-03 08:23
初级用户 Posts 40 Credits 95
Floor29 johnsonlam Posted 2007-08-04 12:24
银牌会员 Posts 497 Credits 1,410 From 九龍,香港
Originally posted by sunx at 2007-7-26 10:56 AM:
DOS programming is very familiar, belonging to the earliest generation of writing programs with debug.com
But I haven't touched DOS for many years,
So before starting work, I once posted a thread here, wanting experienced people to recommend or introduce a few relatively mature DOS development tools now


Not everyone has time to browse every day


I think the optimal combination is DJGPP + ALEGRO
The reasons are as follows:
DJGPP is 32-bit, those who have experience all know why 32-bit code is easier to write than 16-bit code, this doesn't need to be explained more
DJGPP is based on GCC, with guaranteed quality, no need to waste time getting familiar with grammar characteristics
DJGPP is free, no potential copyright risks


DJGPP is not all-powerful either. Its advantage is just being free and having DOS Externder support. The development team is not large, and it is difficult to avoid having bugs...

However, I have used many software developed with DJGPP, such as the Raine emulator, and the compatibility is not bad, maybe because the DJGPP syntax is not "friendly".


This kind of suite is not something that can be well done by 1 or 2 people in 1 or 2 years of part-time writing


This is OpenSource. If you find a problem, you go to report it, and wait for capable people to improve it... That's why you don't need to pay.

It doesn't matter if you don't like DJGPP, there is also OpenWatcom!
If you can use assembly language, NASM or MASM are good choices


Using these things that haven't been strictly tested, don't have special maintenance, lack technical support, and lack practical applications is equivalent to letting the product commit suicide


What you said makes sense. Take a few thousand yuan OS for example:

Microsoft Vista!!

Strictly tested (quite a lot of bugs)
Special maintenance (costs astronomical figures)
Technical support (developers are too lazy to care about you)
Lack of practical application things (still not using things from Win2K, XP!)
Floor30 windowsvesta Posted 2007-08-04 12:43
初级用户 Posts 67 Credits 138 From 云南
The day before yesterday I downloaded "DJGPP" to learn programming, but the compilation speed of RHIDE in DJGPP is too slow, slower than Visual C++, which is unbearable. Now I'm starting to have second thoughts and think it's better to pick up "BC" again.

[ Last edited by windowsvesta on 2007-8-4 at 12:47 PM ]
Prev  1 2 3  Next
[ Contact the Union admin team - 中国DOS联盟 - Standard version ]
Sponsored by ifanr Inc | © 2001–2023