Talk:MS-DOS/Archive 1

Latest comment: 14 years ago by Wendy.krieger in topic Windows NT/2000/XP COMMAND.COM Version (Bug)
Archive 1

Company Name

I think the company name should be Seattle Computer Products, not Seattle Computer Systems. —The preceding unsigned comment was added by 218.223.151.14 (talkcontribs) 22:32, August 6, 2004 (UTC)

Indeed it is, SCP was relatively widely known back in the day for their add-on boards. But a thing that really get on my wits is the "Dirty" part of the Q-DOS acronym. Someone really ought to come up with the goods on this one. Never, never-ever was anything called "Dirty" back in the day. It was known variously as Seattle-DOS, 86-DOS or whatever, but never, never ever as the Quick And Dirty OS. Never ever. Tim (Paterson) himself never called it Quick and Dirty. Never. It's a name that's been slapped on the Q-DOS thingie later on. Quick and Dirty wasn't even really used as a term back in the day, it's a thing of the later 80's. I wish someone could provide some reference to this, since it's driving me up the wall. I keep on reading it here and there, but I never ever ever heard it back in the day. Please please. -Anonymous And Accountless —The preceding unsigned comment was added by 128.214.133.22 (talkcontribs) 02:27, June 22, 2006 (UTC)
Anonymous is incorrect about the name QDOS. The reason he "keeps reading it" is that Paterson himself called it QDOS. Here is the reference he asks for (from Paterson's web site): click here. --Blainster 20:03, 22 June 2006 (UTC)

Release Timeline

I would add this timeline to the article, but it's so sketchy that I'd like to rough it out here first.

Eh...nevermind. I found a good source - will put on main page RobLa 21:33, 18 Sep 2004 (UTC)

Use of backslash

The article says: Fearing copyright infringment complaints from AT&T, Microsoft decided to use backslashes as pathname separators rather than normal forward-slashes. I always thought backslashes were used as pathname separators because forward slashes were commonly used as the switch character in CP/M. Perhaps both reasons contributed to the decision; is it worth adding this to the article?-gadfium 22:42, 13 Feb 2005 (UTC)

In the absence of evidence or a coherent legal theory for the former reason, I've replaced it with the latter and made the sentence indefinite. - User:66.15.40.17 (19 June 2005)
The reason as I understood is is because MS-DOS 1.x doesn't support path names (no directories), so / was used as a switch character rather than a pathname seperator as on other systems (I'd not heard of the CP/M theory but that makes sense, bearing in mind MS-DOS aped CP/M in all sorts of ways). Thus when MS-DOS 2 was released, with pathname support, something else had to be used instead - the backslash was chosen. - User:Retron (14 June 2007)

First sentence

The first sentence says "Microsoft's disk operating system, MS-DOS was the first operating system for the IBM PC."

But the PC-DOS article says: "It is commonplace to see people write "MS-DOS was the operating system for the original IBM PC", for example - which is quite wrong: the IBM PC always shipped with PC-DOS in the traditional IBM blue wrapper."

So this factual conflict needs to be resolved. --Blackcats 00:30, 24 Mar 2005 (UTC)
PC-DOS, as originally shipped, was pretty much MS-DOS with IBM branding. AlistairMcMillan 01:49, 24 Mar 2005 (UTC)
The IBM PC was first introduced with three optional OS's. It was not sold as a complete packaged unit; you had to select the components and software you wanted. But PCDOS was only $60 while CPM-86 cost $240, and UCSD p-System was even more, which accounted for the difference in sales. Blainster 05:10, 20 Jun 2005 (UTC)

Did someone deface this page?

A chunk of it has been removed and there's something in spanish.

Yes... 03:43, 22 May 2006 (UTC) Alexzero77

MS-DOS 'Dominance' Questioned

I would rather it said that it was complimentary to IBM's dominant PC-DOS version. There was just so much code flip flopping how could MS be considered the leader here?

See #First sentence, just half a page above, for details. tyomitch 17:38, 9 November 2005 (UTC)

Expand OEM

Hoping that someone would expand OEM or suitably link it to appropriate articles. Arun T Jayapal 16:58, 22 March 2006 (UTC)

OEM = "Origional Equipment Manufacturer", refers to software shipped with hardware. OEM operating systens come usually with a new computer. One finds OEM versions of NERO (a cd-rom mastering program), that comes with replacement CD-W drives. Wendy.krieger 07:48, 30 August 2007 (UTC)

Who embeds what?

In the article it says several times "Shipped embedded in Windows …", however it was Windows that was embedded in DOS, not the other way around. Up until Windows Me, DOS was still the real OS and Windows just as much a GUI as Windows 1.0 was. Mütze 22:02, 25 April 2006 (UTC)

Uh.... no, that's not quite right. SchmuckyTheCat 22:38, 25 April 2006 (UTC)
I believe that Windows 95 and later were protected-mode 32 bit operating systems independent of DOS where DOS 7 was the real-mode DOS that ran when you selected "MS-DOS Mode". 71.159.225.121 (talk) 23:18, 3 March 2008 (UTC)
One has to be careful here. The Windows 9x package contains comingled code for DOS and Windows. One can quite easily tease out the DOS from some version of Win9x, such as WinME, and replace the corresponding files in an earlier version. With some patching (to activate deactivated code), you can even run Windows 3.1 on DOS 7.1.
Windows 9x requires DOS to boot, and io.sys continues to service some functions under windows once it is loaded. However, much of the traditional DOS interface is handled in protected mode by VxD's. However, it is still possible to exit windows 9x and return to a DOS promot, although this is largely hidden, and the exact process must be unhidden.
One notes too, that something like QEMM services everything in protected mode too, and so "could" be counted as a 32-bit protected mode DOS. Likewise, if one fails to observe that Windows 2.x or 3.x are also operating systems, then how can one count Windows 9x or Windows NT. Not one of these operating systems have a Win32 boot. Every one of these have a virtual machine engine that is implemented in something other than Win32. In this way, they differ from OS/2 and Linux, where even the character boot mode is the same as the operating system presented in the GUI.
DOS is still running when Windows is loaded, to varying degrees. It never goes away. One of the resource limitations that come from Windows 9x is because of the free dos memory limits the number of running processes. --Wendy.krieger (talk) 07:41, 4 March 2008 (UTC)
Besides being able to "exit Windows 9x and return to a DOS prompt", it's also possible to boot Windows 9x to a DOS prompt, then fire up Windows the old-fashioned way by entering "Win". Simply edit MS-DOS.SYS and under "Options" include the line "BootGUI=0".
Booting to DOS in this way allows you to perform all DOS functions. To me, this would seem to indicate that Windows 9x runs on top of DOS, and that the "embedded" DOS is indeed fully functional and totally independent of Windows. Ian Fieggen (talk) 10:39, 4 March 2008 (UTC)
Not really. The Windows Kernel overtakes DOS. DOS is used for compatibility but Windows is not running 'on top' of DOS. DOS is more like a boot loader, just that it's also used for compatibility with real mode device drivers and applications. Assuming you do not use any DOS apps or drivers you should in theory be able to overwrite the DOS memory area with DEADBEEF without Windows noticing. --Anss123 (talk) 12:03, 4 March 2008 (UTC)
I guess we could say that Windows 9x was about the turning point between DOS as the OS and Windows as the OS. Either could operate fully as the OS. If you booted into non-GUI mode, you had DOS as the fully functional 16-bit OS. This was important for certain applications that really didn't like to run under Windows or if you were only running DOS applications and your PC really didn't have the resouces to run Windows' GUI (eg. many POS terminals that I worked on at that time). If you subsequently fired up Windows (or if you booted directly into Windows), it took over all the OS functions and became a full 32-bit OS.
To answer the original question: "Who embeds what?", perhaps the answer should be that the two were a complemetary package. Ian Fieggen (talk) 23:52, 4 March 2008 (UTC)
Quote "But some of the improvements involved reknotting some of the ties with DOS that were too hastily broken in WfW 3.11. For example, recall from Chapter 8 (the "global and local INT 21h calls to local hookers" section) that Wfw 3.11 wouldn't reflect file INT 21h to local hookers, but that Windows 95 will.". Andrew Schulman "Unauthorised Windows 95, Developer's Resource Kit".
Windows, even from version 2, progressively took over handling some of the DOS functions into protected mode. In any form, it is a comingled repackaging of two products, where parts of DOS ship with Windows, and parts of Windows ship with DOS (even in DOS 2.x days). One notes that windows is a DOS32 application, it is possible to extract DOS32 and run other programs in that environment (eg DOSSHELL 5.0 uses a run-time Windows 3.0 loader (dosx)). This version of DOS32 is intended to run just one program, which when one exits, one exits to DOS.
Depending on which version and mode of Windows in installed, gives the name of the single program that DOS32 runs. Both Unauthorised DOS and Unauthorised Windows 95 go into lots of detail on how to convince COMMAND.COM to be this program.
None the same, Windows 95 is an upgrade to Wfw 3.11 + Win32s. Early beta versions had Win32c, for example. The shell is pretty much a proggie one can set in shell= in system.ini[Boot]. You could put sol.exe here!
Windows 9x, (all versions), let DOS administer the PSP data, so even running programs like clock.exe and winhelp.exe in Win95, actually calls down to DOS to create the "Program Segment Process" required for the program to start: ie DOS actually launches the programs.
It is nonsence to say that Windows 9x does not require some form of DOS, but then it is something more akin to a DOS-Windows bundle, for which the bundle has been questioned as anticompetitive in court, and for which Microsoft settled out of court with DR-DOS over the matter (ie it was never decided legally if the bundle is anti-competitive).--Wendy.krieger (talk) 07:52, 5 March 2008 (UTC)
The PSP is used to store the command line arguments and some state information and since Win9x use DOS as a command line it's not surprising that there's some integration there. DOS environment variables are also used by Windows programs. My point was that if one don't need DOS compatibility one should be able to remove it from 9x without 9x noticing since Win9x is pretty much a complete OS (32-bit kernel, 32-bit drivers, 32-bit API), however there are dependencies here and there.
Without DOS Win9x will be unable to start applications, or boot at all. But it’s still wrong to say Windows 95 runs ‘on top of DOS’, or that Windows NT runs 'on top of csrss.exe' :) --Anss123 (talk) 08:57, 5 March 2008 (UTC)
You don't need DOS to launch a command line. The command lines are processed by a DOS utility (command.com), which is quite easily replaced (SHELL=...). Windows handles command lines quite happily in 32-bit mode.
It is correct that Windows bypasses a good deal of 'DOS as OS' functionality when it loads: memory management, file access etc. But this is happening even in Windows 3.1 way back in 1992. You have already in win 3.10 things like 32bit file access. This is implemented by regarding the whole file system as a 32-bit file, and opening and closing files directly by access to the hard disk directly.
If one is going to regard Windows 95 as an operating system, one should regard windows 3.1 as one too! In either case, both use the underlying DOS as a helper VxD, and share code with DOS. The reason that DR-DOS 5.x and 6.x pretends to be dos 3.3 is because Windows makes assumptions about DOS. (Even Windows in NT fails the genuine DOS test).
DOS runs completely without windows, but Windows needs DOS (because its kernel is geared to connect to something like DOS). So it is correct to say that "DOS is comingled into Windows", or embedded in windows, because it can be extracted and run by itself.
Windows NT does not boot from a win32 boot block. Programs that run in the minimalist NT boot, such as O&O bluecon, or the cmdcon utility of windows, do not run in a Windows CMD window. It is a kind of thing written by ex-VMS engineers, utterly unrelated to Win32. Apple were considering running the MAC on it! (project Blackbird).
Windows 9x does indeed run "on top of dos" in much the same way that WordPerfect or QEMM does or much of the gaming world. One has, when loaded to bypass the DOS interface and write directly to hardware. In the days before operating systems, one wrote code that directly to hardware, without pretense of file-system etc. One just managed the memory address directly.
So if Windows 9x does not run "on top of dos", does something like King's Quest 4? --Wendy.krieger (talk) 11:17, 5 March 2008 (UTC)
Windows shares some memory structures with DOS and this includes the command line arguments, but what this have to do with replacing the Windows Shell I fail to understand. Yes I know you can start up Notepad instead of explorer when booting Windows, Notepad - however - is still started with a command line argument and can and do make use of "environment variables", both which as DOS related features.
Now is King's Quest an Operation System? There is no hard line definition of what makes an OS, but they generally should have the ability to run something one could call an 'application'. King's Quest does run scripts which can be argued to be something akin to an application. I'm not sure how powerful KQ's scripting language is, but assuming it's turning complete one could make a word processor for KQ. Does that make KQ an OS? Perhaps it does? Generally, however, people tend to attribute the OS to be the Kernel. This naturally leads into the question 'what is a Kernel?' If you think there's a clear definition of that you're in for a disappointment.
Generally we view the Kernel as the pice of code that handles memory management, disk access, networking and other such services. The Win9x kernel does all this and more. It shares some mem structures with DOS and is geared towards DOS compatibility, but it's still the piece of code that controls the machine.
In case of Windows 3.1, many do in fact view that as an OS. As for KQ, I do not believe anyone views that as an OS but by some definition it might indeed be one.--Anss123 (talk) 14:41, 5 March 2008 (UTC)

DOS

Someone changes 'Disk' to 'Data' in the opening paragraph. Just to avoid an argument, according to microsoft it is Disk [1] -Localzuk (talk) 13:39, 3 July 2006 (UTC)

Where'd the code come from?

From the article, "However, MS-DOS 4.0 was actually based on IBM PC-DOS 4.0, as Microsoft was by then concentrating on OS/2 development." Does this mean that IBM actually wrote the code for MS-DOS 4.0 or that they had merely relicensed with minor improvements the code it already had from Microsoft? This is all rather confusing. It would be nice if a chart like Image:Unix.png could be creating for the DOS family of operating systems. Theshibboleth 12:08, 14 July 2006 (UTC)

Most of the new code in PC-DOS 4.0 was written by IBM but the starting point was DOS 3.3. DOS 4 shipped on 5 360k disks; DOS 3.3 only needed 2 360k disks. IBM added a lot. The larger partitions were derived from the code used in Compaq DOS 3.31. DOSSHELL was derived from IBM's Fixed Disk Organizer. Addditional code was created by IBM to support all the hardware introduced as part of the PS/2 line. Expanded memory support would only work on IBM memory boards and PC-DOS 4 would only read hard disks formatted by PC-DOS. 24 OCT 2006


Replacements.

I think a new section at the end of the article called "Replacements" could be a good idea. I've noticed FreeDOS and DOSBox are already mentioned, but for newbies trying to run old MS-DOS programs on other OS it could be more helpful if expressed in this fashion. —The preceding unsigned comment was added by 217.125.97.122 (talk) 10:43, 21 December 2006 (UTC).

Long hiatus between DOS 4.01 and DOS 5.0 - why?

The time between the 4.01 and the 5.0 releases is by far the longest period without a new version, 2.5 years to be exact. Why was there such a long break at that time? Was everbody working on Windows 3.0? -- 85.179.170.121 13:25, 12 May 2007 (UTC)

Much of Microsoft OS development staff was working on OS/2. There was some work on improving Windows but most accounts indicate that team was small.

I joined the MS-DOS support team in 1990 as part of the "build-up" for 5.0 (which wasn't released til 91). I was told that 4.x was largely an IBM effort; Microsoft had shifted people to Windows and OS/2, thinking that customers wouldn't want DOS anymore. That's hearsay, tho - I didn't transfer to the MS-DOS product team until 1992. I can say the MS-DOS 6.x team was smaller than the 5.0, and much smaller than the Windows & Windows NT teams. JenKilmer 16:45, 29 May 2007 (UTC)
On the street, MS-DOS 4.x proved very unpopular. Perhaps this caused them to pause and re-group? It would also undoubtedly have affected sales (and hence revenue).
The best "computer" use I ever saw for MS-DOS 4.0 was two original, unopened, boxed copies being used as a dot matrix printer stand! Ian Fieggen 03:47, 31 May 2007 (UTC)
The problem with comparing 4.x sales with 5 & later is that Microsoft did not package 4 for retail sale. Microsoft only licensed it to OEMs for sale with original machines. (Yes, I know some copies still made it into the retail channel - not the point.) And OEMs could decide whether to ship 4 or 3.3. So yes, 4.x was not only larger and without much obvious benefit, but relatively rare on the street...not a good way to create a good rep! JenKilmer 18:23, 31 May 2007 (UTC)
Additionally the DOS 4.x series not only brought few benefits but it did bring many problems. I forget exactly what the problems were now and exactly which versions were affected but I do recall that until version 5.0 was released it was necessary to stick with 3.3 in order to have a useable system. —Preceding unsigned comment added by 213.162.108.220 (talk) 23:15, 21 September 2007 (UTC)
DOS 4.x introduced ideas that were eventually passed onto OS/2, such as installable file systems. It had better NLS support, and support for such as ramdisks, extended memory, and large disk support [fat16], although fat16 appeared in DOS 3.31. Dosshell appeared here too. There was too much new stuff in it, and unfortunately, too many bugs. --Wendy.krieger (talk) 11:36, 25 November 2007 (UTC)

From the section "Legacy compatibility"

This is from the section "Legacy compatibility": "From 1983 onwards various companies have worked on graphical user interfaces (GUIs) capable of running on PC hardware. With DOS being the dominant operating system several companies released alternate shells, e.g. Microsoft Word for DOS, XTree, and the Norton Shell."

How is Microsoft Word for DOS an "alternate shell" similar to XTree or Norton Shell? What is possibly meant here? 131.111.223.43 19:03, 2 June 2007 (UTC)

You got me. Word is an application. When I think alternate shell for DOS I think of a command.com replacement. Perhaps they meant User Interface toolkit.
--Anss123 19:37, 2 June 2007 (UTC)

Merger of DOS and MS-DOS

Missing PC-DOS versions

Missing from the list are these PC-DOS versions
After the IBM/Microsoft breakup, IBM decided to have a go at the retail market. Generally, IBM was active in this market between 1992 and 1996 (when Bocca Raton closed). OS/2 4.00 was the last real release from here, DOS 7,00a and OS/2 4.50 are just slipstreamed bugfixes. Later PC-DOS 7.10 is built of a segment of the code, largely for building machines, although there is a version in Norton Ghost. So far, four different versions of 7.1 have been collected.
  • IBM-DOS 5.02 date 09-01-1992 First dos to include INTERLNK.
  • IBM-DOS 6.10 date 06-29-1993 Includes support for DBLSPACE compression, but none included
Microsoft did not license DoubleSpace to IBM. IBM licensed Stacker for use in PC-DOS 6.10. In the interest of full disclosure, I worked on the MS-DOS test team for Microsoft during the MS-DOS 6 releases. JenKilmer 07:02, 30 August 2007 (UTC)
IBM-DOS 6.10 was released with no disk compression, although it will load DBLSPACE in the manner that MS-DOS 6.00 does. PC-DOS 6.10 and 6.30 includes Superstor, loaded in the same way that MS-DOS loads dblspace. PC-DOS 7.0 and 2000 will load the first of Stacker.bin or dblspace.bin. Wendy.krieger 07:14, 30 August 2007 (UTC)
  • PC-DOS 6.10 dos 11-16-1993 Includes Superstore compression.

Wendy.krieger 08:49, 23 August 2007 (UTC)

DOS versions

Of these versions:

  • MS-DOS - July 1981 (July 1980 as QDOS) this is one that most people know
  • PC-DOS - August 1981. IBM recompiled the source using alternate compilers
  • DR-DOS - 1988 = first retail DOS, from ver 3.41
  • Novell DOS - 1991 = DR-DOS after novell bought it.
  • OpenDOS - DR-DOS source, when Caldera owned it.
  • FreeDOS - Open source attempt to recreate DOS.
  • ROM-DOS - Original source: Datalight makes embedded systems.
  • PTS-DOS - Russian version, to prevent external sabotage.
  • GS-DOS - General software DOS = another embedded system

Wendy.krieger 08:57, 23 August 2007 (UTC)

Split_section_out_proposal

AND...

{{Merge}} See the talk page at Talk:DOS 130.101.20.153 16:23, 4 December 2007 (UTC)

Made this into the common discussion page on the proposal and fixed all the links to point here. The nom didn't make much of a case, but there's some merit to removing the data from the articles.

In addition to the split out, this discussion has the co-proposal of merging the three sections of mainly redundant material into one list article or perhaps a template page
  1. http://en.wikipedia.org/wiki/PC-DOS#Versions
  2. http://en.wikipedia.org/wiki/DOS#Timeline http://en.wikipedia.org/wiki/PC-DOS#Versions
  3. http://en.wikipedia.org/wiki/MS-DOS#Versions and release dates
Support -- To my way of thinking, a template using a table format would be one solution, for it could clean up the presentation, and such.

2. Alternatively, making the page into a comparative development and release history list article with some comments would unclutter the current articles and offer the benefits of having one page that had definitive history, as now an fixup edit or cite can be made in one, but not be seen in the others.

2.1. It would however, leave the PC-DOS article short of material, comparatively, but as I hold IBM in dispite, no matter what, I can live with it.

3. Third option: Do the tables up as a common template, if some material is to be excluded, use a page test like:

{{#:ifeq:{{PAGENAME}}|MS-DOS|<includeonly>
...lines to be included only for MS-DOS page
</includeonly>
}}

{{#if: {{ifequal|{{PAGENAME}}|MS-DOS|{{PAGENAME}}|PC_DOS}}|<includeonly>
...
...lines to be included both for MS-DOS and PC-DOS page
...
</includeonly>
}}
3.1. This would be a much better option in a template solution, as people familiar with the coding would tend to be watching it.
3.2. Oh, if template implemented, I'd suggest looking at doing it up as a sidebar... See the historical sidebars in BASIC, but this one could be a table format--would give a better appearance I'd think. Ask if you all want help. I've my normal beat to get back to. Cheers! // FrankB 05:36, 10 December 2007 (UTC)
Comment This should be named "Timeline of DOS", as it would contain PC-DOS and DR-DOS as well. 130.101.163.216 (talk) 15:38, 13 December 2007 (UTC)
Never mind, I went ahead and changed the template. Seeing the Afd discussion for DOS suggesting it should only redirect to MS-DOS, I think consensus is that the generic term be used.
As for the above options, support a separate page in list format. I'd worry about a template, because it would either A) be too long to be put in MS, PC, DR articles without them becoming too long, or B) omit a great deal of info that's there already. As for splitting it with the includeonly tags, it loses a lot of its purpose if the material on all of the operating systems isn't together. Plus, on a separate page something like a graphical timetable could be done.
For the PC-DOS article, maybe a little more on the subtle differences with MS-DOS, and a lot more on the post split versions (later 6.x and PC-DOS 7) 130.101.160.206 (talk) 19:44, 13 December 2007 (UTC)
There is a pretty long timeline at Comparison of x86 DOS operating_systems, how about removing the mentioned timelines and pointing to that page? 69.221.152.25 (talk) 19:09, 3 January 2008 (UTC)
I just edited DOS to link there, maybe this will be a good fix. 69.221.152.25 (talk) 23:12, 3 January 2008 (UTC)
MS-DOS, PC-DOS and DR-DOS all have a common ancestor and their histories intertwine, although DR-DOS had an earlier split. One could have a common "DOS Timeline", but i would not be tempted to merge the complete topics under DOS. On the other hand, MS and PC DOS do share common links to 5.00, but beginning with 5.00.1, they diverge under different managements.
Under MS-DOS I would include DOS sessions from Windows NT, since this is a hacked version of MS-DOS 5.0 (mostly by removing things). The DOS that appeared with both early Win95 betas, and with Windows NT are DOS 5.0, even down to including QBASIC 1.0, rather than 1.1 of MS-DOS 6.x. One of the early Chicago betas (Win95), even included an option to use MS-DOS 6.x QBASIC.
MS-DOS 7.x uses a different boot sector, and can co-boot with earlier DOS versions. One can replace the DOS on some Windows with a DOS from a later version, if all of the DOS files are replaced. As with earlier DOS versions, we still have the standard drivers being version-free (eg himem.sys, emm386.exe, ramdisk.sys, mscdex.exe, smartdrv.exe), all work under older and different DOS versions.
Under PC-DOS I would include OS/2's DOS, for 1.x = DOS 3.3, under 2.x+, 5.0, and under OS/2 PPC, 6.3. IBM DOS versions are rather more interesting, because their releases show some sort of increasing feature set: OS/2 PPC version of 6.3 includes REXX in the same style as PC-DOS 7.0, but because of the way it is linked to OS/2, the files do not run correctly under PC-DOS 6.3 for i386 chips. The most recent is PC-DOS 7.1, used to access large disks with fat32 formatting. It is a recompile of parts of the moribound DOS (after Boca Raton closed). --Wendy.krieger (talk) 09:42, 4 February 2008 (UTC)

Extended VER Text

I've gotten a few strings out from the Windows XP Professional CMD.exe frontend.

Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:\Documents and Settings\Benjamin>command.com /k ver

MS-DOS Version 5.00.500

C:\DOCUME~1\BENJAMIN>command.com /?
Starts a new instance of the MS-DOS command interpreter.

COMMAND [[drive:]path] [device] [/E:nnnnn] [/P] [/C string] [/MSG]

 [drive:]path    Specifies the directory containing COMMAND.COM file.
 device          Specifies the device to use for command input and output.
 /E:nnnnn        Sets the initial environment size to nnnnn bytes.
 /P              Makes the new command interpreter permanent (can't exit).
 /C string       Carries out the command specified by string, and then stops.
 /MSG            Specifies that all error messages be stored in memory. You
                 need to specify /P with this switch.

C:\DOCUME~1\BENJAMIN>command
Microsoft(R) Windows DOS
(C)Copyright Microsoft Corp 1990-2001.

C:\DOCUME~1\BENJAMIN>
  • What DOS says it is, and what something like "truever" returns, is two different things. NT's DOS is DOS 5.0, pretending to be build 5.5. While both commands come from command (because you get the same result from 4nt.exe), the second is for DOS compatibility. Only the second environment supports TSRs at the prompt.
  • Judging from the version number 5.00.500, DOS has not changed from Windows NT 3.1. —Preceding unsigned comment added by Wendy.krieger (talkcontribs) 05:42, 28 January 2008 (UTC)

Would this image benefit the article?

Captured it myself in MS-Virtual PC 2007. MS-DOS 7.10 Startup splash Maiq the liar (talk) 16:09, 16 December 2007 (UTC)

I would say probably not. The existence of an MS-DOS 7.1 contradicted the article, and I dug around to see what is actually is. It appears that someone merely took a Win 98 boot disk and tacked on the old DOS utilities that had been packaged with Windows, and are now making it available for download. It's not done by Microsoft, and I doubt MS would be happy if they found out about it. See these links:
69.221.142.252 (talk) 16:41, 17 December 2007 (UTC)

Wouldn't that be a windows '95 boot disc? i though 98 used 8.0?...Aww, i thought i contributed something useful... Maiq the liar (talk) 00:20, 26 December 2007 (UTC)

95 and 98 used 7 as their DOS version number, DOS 8 was under Win ME and the DOS pieces stuck into the NT based systems. 69.221.152.112 (talk) 20:15, 27 December 2007 (UTC)
The DOS peices in Windows NT are 5,0, hacked to say 5.5. It's only when you ask it to make a DOS diskette that it creates DOS 8.03 (See STARMAN: mirror.href.com/thestarman/winxp/winxpsd.htm This says that it's a hacked version of the Windows ME disk, with files deleted etc).--Wendy.krieger (talk) 10:42, 4 February 2008 (UTC)

"This article is about the family of closely related operating systems for the IBM PC compatible platform"

then it should be where it belongs, on the MS-DOS page. DOS is a generic acronym standing for Disk Operating System. —Preceding unsigned comment added by A plague of rainbows (talkcontribs) 16:25, 28 December 2007 (UTC)

No, MS-DOS only refers to the Microsoft's version and by extension the other versions based on it. DOS is a generic name for operating systems that copy MS-DOS, as well as an acronym for disk operating system. - Josh (talk | contribs) 17:23, 28 December 2007 (UTC)
Why do operating systems that copy MS-DOS merit a page of their own? How many people ever used them relative to MS-DOS itself? A section on the MS-DOS page would be better, especially in light of all the other DOS's that have nothing to do with MS-DOS.
Good point. It could be like the Emulation software section of Microsoft Windows. - Josh (talk | contribs) 18:48, 28 December 2007 (UTC)
MS-DOS is one of many operating systems that use a "DOS-like" interface. One can speak of such, compared with the UNIX/Linux paradigm, in that DOS use single-letter block-drivers (C:, D:, E: &c), backslash path separators, EXTPROC, etc, in much the same way that Unix uses a single file system (with all devices mounted in the root file system), forward slashes, multiple shells, and The /# command for IBM's extproc. Operating systems like OS/2, and Windows, Windows NT, are clearly not DOS, but are "DOS-like".
The history of DOS is tangled in the Digital Research/Microsoft/IBM stuff, but once it became widely available, other vendors implement clones of the API to simplify learning + get their own licence fees for the product. None the same, different DOS versions have competed (eg MS-DOS 5, 6, Win95, are largely responsenses to things like DR-DOS, PC-DOS 6.3, and OS/2. NT was not ready for prime time at that stage. --Wendy.krieger (talk) 11:11, 9 February 2008 (UTC)
Props on the "unix-like" term comparison. Some of the page move vandalism has placed DOS at unnecessarily long article names with the purpose of putting that "MS" in the title. I disagree with doing that for a few reasons: 1) "DOS" is much, much shorter; 2) This is the most common usage of "DOS" as compared with the other meanings at DOS (disambiguation), and was used in a number of DOS programs (as opposed to referring to the OS as "MS-DOS", which they did only occasionally); 3) That's giving MS-DOS sole credit as though it was an unprecedented technological breakthrough- it wasn't. First, 86-DOS was the original name, so something like "86-DOS Compatible Operating Systems" would be more appropriate. Second, it was very nearly a clone of CP/M. The binaries weren't compatible, but the interface, drive letter assignment, etc. was identical, and no other systems before CP/M had any of these. Tim Paterson even said it was designed to operate the same (see the quote at Tim Paterson). If we call DR-DOS and FreeDOS "MS-DOS Compatible Operating Systems", then we should call MS-DOS a "CP/M-like Operating System". I'm not advocating that, but I do believe that DOS is an OS family with too jumbled of a beginning to place it on the shoulders of Microsoft. JeremyMcCracken (talk) (contribs) 22:49, 7 April 2008 (UTC)
DOS and Unix are the two most common paradigms for command-line user interfaces. Neither of these could have been said to invent it, but were the main users of these paradigms when computers came into the public attention. By the time this happened, there were already several Unixen and several DOSes in the market place. I think it is appropriate to recognise that DOS is an interface idiom shared with several operating systems that mostly derive from CP/M (eg DOS, OS/2, Windows), while there are others that derive from the Multix/Unix interface (eg Unix, BeOS, Apple OS/X). The x86 reference is probably not important, since Microsoft had NT running on Risc chips (PPC, Alpha, Mips), and there is an OS/2 for PPC as well. I really think the interface is something to do with IBM, rather than Microsoft, in any case, since they required PC-DOS to have a CP/M interface. --Wendy.krieger (talk) 10:10, 8 April 2008 (UTC)

Origins

imho a starting section with origins or the like would be useful. as far as i remember, the first ms-dos started as a copy (or better, the code was acquired) by Bill Gates, wo then entered into an agreement with IBM to pre-ship the then to be release first ibm pcs. if anyone agrees and feels more comfortable in starting this section. otherwise i may give it a try. thanks. --BBird (talk) 16:55, 4 February 2008 (UTC)

Bootfloppy?

Gah, how do i go about making a MS-DOS Bootfloppy if i don't Have a floppy drive?! Maiq the liar (talk) 13:18, 7 February 2008 (UTC)

You go to bootdisk.com --Anss123 (talk) 15:18, 7 February 2008 (UTC)
WinImage http://www.winimage.com/ will also make a boot floppy for both Windows 9x and MS-DOS. You need to have the appropriate files available though. Alternately, you can use bfi.exe (boot floppy image), from Bart Largerej http://www.nu2.nu/ where you supply the boot sector code (eg from another diskette), along with the diskette layout. You can order up to ten files on the diskette. In either case, such diskettes can be used as boot images, or copied (eg by IBM's loaddskf or GNU rawrite to a real floppy diskette elsewhere.--Wendy.krieger (talk) 11:00, 9 February 2008 (UTC)

Merging DOS article here

The old discussion appears to be about the reverse. As the current DOS article talks entirely about MS-DOS-derived operating systems, I think it'd be more appropriate as a section here, and have DOS be a redirect to Disk Operating System since it's a pretty common acronym.Petchboo (talk) 19:38, 7 March 2008 (UTC)

This has been suggested before, and the consensus was to keep them as they are. See Merger of DOS and MS-DOS above and the deletion discussion. 91.121.64.135 (talk) 01:08, 13 March 2008 (UTC)
Links for Incident of a user moving DOS into MS-DOS with no consensus which got reverted because he didn't discuss it before, and people oppossed it anyways --Enric Naval (talk) 09:03, 14 March 2008 (UTC)

History section

I tagged this as delete. It's written well, but is basically a duplicate of the history section of DOS. There had been concerns before of MS-DOS, PC-DOS, DR-DOS, and DOS all being too long because they all contained nearly identical information. Though MS-DOS was derived from the earliest DOS (86-DOS), DR-DOS and DOS Plus are tied to CP/M through DR, so it seems best to keep the history there. 91.121.64.135 (talk) 01:23, 13 March 2008 (UTC)

I agree with merging the history section with the DOS article --Enric Naval (talk) 07:59, 13 March 2008 (UTC)
As the contributor of the History section, in principle I agree with merging; I wasn't aware of the duplication. Ideally there should be one main article, which if we follow the history of the family should be MS-DOS (originally a family of operating systems for 808x-based machines of any architecture), with PC-DOS (MS-DOS for PC, licenced by IBM from Microsoft), and the much later DR-DOS, Novell DOS, FreeDOS and all the others deriving from it.

One point that I think is important in an article on MS-DOS in particular is that it was designed to be machine-dependent, the interface between standardised programs and disparate hardware; it only became identified with the PC and clones when they wiped out all other hardware. Many people are totally unaware of that (if they have heard of MS-DOS at all...)

And plain DOS is a mainframe operating system, called simply DOS (only later DOS/360); there has never been a microcomputer operating system called "DOS" to my knowledge. A simple disambiguation of DOS to DOS/360 or MS-DOS is called for.Pol098 (talk) 22:32, 13 March 2008 (UTC)
You're right, the DOS article is actually talking about "MS-DOS" everytime it mentions "DOS" anywhere on the article. Merging DOS's contents into this article and then making a disambiguation would be adequate. The disambiguation could have also a list of DOS systems, like the one on the "related systems" section on this article There already exists a disambiguation page for DOS --Enric Naval (talk) 08:49, 14 March 2008 (UTC)
Ok, I see there has been a lot of trouble with that proposal, and that we simply won't be able to merge the DOS article because so many people opposse for several reasons (mostly anti-Microsoft reasons, imho, but, oh well, whatever).
How about this? The DOS article covers MS-DOS history in excessive detail, covering facts that don't help to understand what a DOS is. So, let's move only the bits specific to MS-DOS from the DOS article to here until there is only the minimum duplication necessary. This would address the duplication problems pointed out by Pol098 and even improve both the DOS and the MS-DOS articles --Enric Naval (talk) 09:11, 14 March 2008 (UTC)
That entire article is MS-DOS specific; that's my point. DOS and Windows, versions, accessing hardware under (MS)DOS, other PC operating systems, reserved device names under (MS)DOS, drive naming scheme, (MS)DOS boot sequence, DOS emulators... what could be left? The entire thing needs to be merged or renamed, or reverted - it started out as a general DOS article but got co-opted by MSDOSites
I think the purpose of having DOS separate from MS-DOS is it includes the Digital Research involvement- they were in the mix when MS-DOS emerged, and later DR made DOS Plus and DR-DOS. To make it an MS-DOS article would inherently push this information out to cover only the Microsoft development. As I look at the history section of the current revision of DOS, I see two paragraphs that detail the development of 86-DOS and the DR-IBM fallout, one about MS/PC-DOS, one about DR-DOS, and one about FreeDOS. It seems pretty even with that configuration. JeremyMcCracken (talk) 05:13, 25 March 2008 (UTC)

I just trimmed "History" down; I think this keeps some MS specific information without being a duplicate. JeremyMcCracken (talk) 11:17, 25 March 2008 (UTC)

sync problems

See what happens when we have duplicate information on two articles? Now this info needs getting copied here [2] --Enric Naval (talk) 13:00, 2 April 2008 (UTC)


Dreadful article

I think the english in this article is truly dreadful and the whole thing should be rewritten. —Preceding unsigned comment added by 82.23.145.183 (talkcontribs) 14:51, 7 April 2008 --Enric Naval (talk) 15:05, 7 April 2008 (UTC)

Wikipedia can be edited by everyone, so you can do it yourself --Enric Naval (talk) 15:05, 7 April 2008 (UTC)

Imbalanced Text

"The (once infamous) AARD code, a block of code in the Windows 3.1 beta installer. It was XOR encrypted, self-modifying, and deliberately obfuscated, using various undocumented DOS structures and functions to determine whether or not Windows really was running on MS-DOS. [2]" "Interrupt routines called by Windows to inform MS-DOS that Windows is starting/exiting, information that MS-DOS retained in an IN_WINDOWS flag, in spite of the fact that MS-DOS and Windows were supposed to be two separate products. [2]"

These two items can have perfectly innocent explanations, but this is not mentioned. Why would Microsoft want to complicate the Beta test process by allowing it to run on any DOS platform? In case 2, MS-DOS may need to know when it is running in an MS-DOS prompt rather than by itself. We should be careful that concerns of alleged guilt in some areas should not be causing presumption of guilt in others, as that is not how the world really works. Programmers quite often do things to acheive a technically desirable result, such as improved performance: not every line of code is vetted by an alleged committee of corruption and dirty tricks. —Preceding unsigned comment added by 124.168.211.6 (talk) 23:56, 1 May 2008 (UTC)

  • The AARD code appeared in HIMEM.SYS (see http://members.ozemail.com.au/~geoffch@ozemail.com.au/editorial/aard/index.html ). The code appears in a number of utilities, including win.cnf (a part of win.com). We note first that it is XORed and encrypted etc. This suggests that its purpose is not meant to be public. The particular tests were to see if Windows were running on MS-DOS or PC-DOS.
  • The actual test is for something that is not used by Windows. Further, Windows makes less assumptions about pre-dos 5 versions, and would quite happily run on DR-DOS 5.0 (which pretends to be 3,31). The tests are in effect an attempt to use undocumented DOS to sniff out DR-DOS.
  • Microsoft is indeed motivated to do this sort of thing, and this exists with precident. One can, in early versions of PWB (the editor from several DOS compilers), get a message of the effect (PWB help message .9050 "Non-Microsoft Operating System"
    • The Microsoft Programmer's WorkBench (PWB) has found that the operating system is not MS-DOS or PC-DOS; that is, it was not originated by Microsoft. If PWB is run on an operating system that is not an exact duplicate of a Microsoft-originated operating system, the product may not perform as intended.
  • The versions that bring this up in DR-DOS will do this under Windows NT as well.
  • It's important to realise that Betas are distributed to large numbers of swavvy techie types, including media people. Such feuls magazine articles. To have the message come out every time that DR-DOS starts a Windows beta, will increase the uncertianty that DR-DOS has enough differences not to run Windows.
  • What was driving DOS 5, DOS 6, DOS 7, is that DR-DOS was innovating DOS features in advance of Windows (and Microsoft was effectively cloning DR-DOS). On the other hand, Microsoft could comingle things into Windows that would leave DR-DOS a new application that DOS must run. So Microsoft was trailing always on the feature set, but could dink the required API through Windows. --Wendy.krieger (talk) 10:06, 4 May 2008 (UTC)

Misplaced comments in History Section

The last paragrapg in the "History" section needs to be removed, moved or revised. The last section is mostly a criticism of IBM dependancy at that point in time, with little focus on MS-DOS. I suggest moving this criticism to an appropriate IBM-Computer page. Furthermore, the information that is relevant to MS-DOS should be reworded in a way that does not focus on IBM. I am not knowledgable enough to write the actual content. Thanks. TheTrueHeadfoot (talk) 03:09, 14 May 2008 (UTC)

System operating on MS-DOS?!

How the heck can the old Windows products be based on MS-DOS? MS-DOS is a command prompt program, not the computer itself! HyperSonicBoom 00:34, 8 February 2007 (UTC)

These old Windows products are based on MS-DOS because they ran from DOS base. DOS in turn is based on BIOS.83.5.24.189 22:12, 18 March 2007 (UTC)
Finally BIOS is based on hardware itself.83.19.52.107 07:33, 19 March 2007 (UTC)
Windows (any version), requires an environment created by a pre-existing loader. Under Windows 1.x-3.x, this was done by an external version of DOS (MS-DOS, PC-DOS, DR-DOS). Beginning with Windows 9x, this is done by an inbuilt version of DOS, eg MS-DOS 7.x or 8.x. It is possible to replace this DOS with one that has at least the same version number, as has been shown with DR-DOS and Windows 95a. When Windows is loaded, it is used as a virtual mode driver, to varying less degrees.
Windows NT (all versions), rely on a different system, written by programmers from the VM world, but made to look like OS/2. In any case, none of these versions will support even the cmd.exe program until all of it is loaded.
Command.com is not "MS-DOS", or even a part of the operating system. It's a bundled application that can be easily changed by redirecting the SHELL= line in config.sys. The popular shareware program "4dos" does exactly this. Also, even if the program is called 'command.com', it does not need to result in a c:\> prompt. The IBM system diskettes were basically PC-DOS 4.01 with a modified IBMDOS.COM and an extensively modified COMMAND.COM.
In Windows 3.x and 9x, it is correct to use the MS-DOS icon to show the command prompt. In Windows NT 3.x and 4.x, this is incorrect, but leads to this confusion. --Wendy.krieger (talk) 11:27, 25 November 2007 (UTC)
Older versions of Windows use MS-DOS as preloading system. If you have double-boot with Vista you might have seen that its boot loader is also based on MS-DOS! —Preceding unsigned comment added by 89.172.140.94 (talk) 17:42, 19 October 2008 (UTC)
Vista does not use MS-DOS as a bootloader. SchmuckyTheCat (talk)

Raking over the past

From the back-channels - please ignore the somewhat partisan language, these opinions are backed by independent sources and the opinion expressed is not batshit, there have been TV documentary serials which have made much the same points (in somewhat more measured language). The article already says some of this, but the complainant states with some justification that the balance is still overall too favourable to Gates, who did, after all, lose his intellectual property case with Kildall. Anyway, here it is:

I really have just two concerns. One involves the pretty

clear fact that Bill Gates and IBM stole CP/M from Gary Kildall, that MS-DOS and PC-DOS were not just CP/M clones, but CP/M renamed. Proof for this came out at trial. There were lines of code in MS-DOS that only Gary Kildall could explain. Neither Bill Gates nor any programmer at Microsoft could explain these lines. This resulted in Microsoft settling the case and paying Gary Kildall a lot of money. As is usual in legal settlements, the winning party often has to sign an agreement not discuss the case. This makes Bill Gates the biggest software pirate of all times.

Bill Gates not only stole CP/M from Gary Kildall, but it was proved that he engaged in many illegal practices to destroy DR-DOS. "DR-DOS" was Gary Kildall's clone of later versions of MS-DOS.

There was a legal way to clone a competitors' software and an illegal way. The illegal way involved stealing software code line for line. The legal to clone software involved writing original code to mimic the software being cloned. You can't tell by using software to tell whether it is a legal clone or whether it is using stolen code. Software code has always been protected. But in most cases you can't protect software functionality.

Gary Kildall proved that DR-DOS was a legal clone of later versions of MS-DOS. Gary Kildall did not steal any lines of code from MS-DOS in making DR-DOS. On the other hand, the original MS-DOS was line for line a virtual copy of CP/M.

Bill Gates e-mails to others paint a very unflattering picture of Bill Gates. They show he tried every underhanded trick he could to hurt his competition, especially Gary Kildall who he feared the most. Bill Gates spread lies that software written for MS-DOS would not run on DR-DOS.

Bill Gates knew nothing about operating systems. Gary Kildall was the genius who created DOS for micro-computers. All Bill Gates had done was write BASIC for micro-computers. All Tim Paterson did was illegally modify CP/M to run on Intel's 8088 chip. Seattle Computer Products needed a micro-computer operating system for a new computer they were building using the 8088 chip. Gary Kildall did not want to sell them CP/M. So, Seattle Computer Products stole CP/M. After Gary Kildall refused to modify CP/M for them, they had a one of their programmers do it for them. Tim Paterson was not the kind of programmer who could create an operating system

Bill Gates and Tim Paterson knew that Tim Paterson's Q-DOS was really CP/M when Bill Gates paid Tim Paterson for rights to use it

There is a very good reason Bill Gates did not pay Tim Paterson a lot of money for Q-DOS, and why IBM did not pay Bill Gates a lot of money for PC-DOS. All these players knew they were illegally pirating CP/M.

The American legal system makes it profitable for big companies to steal the products of others. When a big company has been found guilty of having infringed on the patent of another company, or some person all they usually have to do is pay some small royalty. Very rarely does the original patent holder get back their true loss.

It often makes more economic sense to steal a patent and pay the amount determined by a judge years later, then negotiate with a creator and inventor who doesn't want to sell his baby. Large corporations gamble that judges are going to value the worth of a patent lest than the inventor, even when the company who steals the patent has made billions off it. What judge is going to award a billion dollar judgment?

Finally, Bill Gates grossly violated the Sherman Anti-Trust act for many years by threatening computer manufacturers that he would not sell them or allow them to use any Microsoft product if offered customers computers with competitor's products.

I have documentation to prove this from the lawsuit Gary Kildall filed against Bill Gates. All Wikipedia articles that involve this issue give Bill Gate's discredited defense. Gary Kildall's side is either not presented or not presented objectively.

There is a lot of criticism of Bill Gates that has been sanitized. For example, Bill Gates took over second or third rate applications then made them the de facto standard by using his virtual monopoly. He would get computer manufacturers to give away his second-rate applications free on all their computers, until his applications became the industry standard because they were so prevalent.

Most lawsuits against Microsoft have been successful. For example, Apple successfully sued Microsoft for stealing their Interface. Xerox lost their lawsuit against Apple. Apple took the general idea of GUI from Xerox. You can't patent an idea. Apple didn't copy Xerox's GUI, they transformed it into an entirely new product. On the there hand, Bill Gates and Microsoft did not improve Apple's interface. Windows 95 was just an inferior version of Apple's interface. There was no innovation and very little change.

Japan used to take American products and improve them. Bill Gates took competitors' products and didn't improve them. He simply used illegal monopolistic practices to turn his products in the de facto standard.

I wish the true story of Bill Gates was told or there should at least be a section where the public can learn of the extent of the criticism of Bill Gates and his business practices.

Gary Kildall won his case against Bill Gates. Apple won its case against Bill Gates. The Attorney General's of 26 states won their case against Bill Gates. And these just represent the tip of an iceberg. 90% of the cases filed against Microsoft are successful, and this is not because Microsoft settles cases quickly or easily. Microsoft fights every case tooth and nail. It is just the American legal system seems averse to destroying a legend, and a company that is so popular. Microsoft and Bill Gates are considered by many to be an American success story, when the truth is in fact a lot darker.

I can't think of another company that has succeeded by skullduggery the way Microsoft has succeeded.

Microsoft has a lot of intelligent people working for it. Considering the vast talent pool it encompasses it is surprising it has come up with so few original

products.

I have advised the individual concerned to take this up himself, but his past experiences on Wikipedia have been less than pleasant so he may not be inclined to do so. I have at least three too many content disputes on the boil right now, so I hope that someone reasonably sympathetic will at least have a look at this. Thanks, Guy (Help!) 11:14, 1 April 2007 (UTC)

What independent sources? This seems like a run of the mill conspiracy theorist ranting to me. The rant goes on and on about how digital won the case (a settlement is a settlement, not a victory) and MS copied CP/M (twenty years of looking have not turned up any CP/M code in DOS.) It probably touch upon truths here and there, but with it holding statements like: "Tim Paterson was not the kind of programmer who could create an operating system" I somehow doubt it.--Anss123 15:38, 1 April 2007 (UTC)
If it was that cut and dried about who originally created DOS (as according to Kildall) then Paterson wouldn't be suing Kildall's estate (a current case, 2006 or so, I'd have to look it up) for defamation of his character for publishing that version of the facts. For other "facts" like Apple vs. Microsoft our article says "Apple lost all claims in the lawsuit". Microsoft did a lot of underhanded crap, particularly in the late 80s to late 90s.
If an editor wants to come along with notable and attributable edits we can work with them. SchmuckyTheCat 15:45, 1 April 2007 (UTC)
Actually, it's perfectly true that Paterson didn't have the ability to create an OS. But that doesn't prove that he stole CP/M. Quite the opposite. No serious software expert considers QDOS a real OS. It's more like a program loader that pretends to implement various CP/M APIs. Many important OS features (scheduling, I/O libraries) were either missing altogether or so badly implemented as to be useless. That's why IBM clones had to re-implement the BIOS so carefully: application programs routinely bypassed MS-DOS and accessed the BIOS directly, using their own implementation of basic features (like simple I/O buffering!) that is normally provided by the OS. § If our anonymous conspiracy theorist submited a lot of unsourced nonsense like this, it's small wonder he feels unwelcome at WP. Isaac R (talk) 17:58, 20 October 2008 (UTC)
Isaac's comments here are confusing here. We must recall here that DOS is even here a 16-bit OS, that would ultimately displace established eight-bit OS's like BASIC, CP/M &c. DOS 1 as it stands, is a kind of 16-bit program-loader that supplied limited API calls (eg file-system). It's not multi-tasking (This appears in DOS 2, which is pre-emptive), nor does it provide I/O queues. It's written for software that ran the machine live, busy-waiting on the I/O devices as needed. It's different to say that "Patterson wrote QDOS", against "Patterson wrote an OS with I/O management, scheduling ...".
Whether something is an OS is a question of doubt. DOS has been described as an 'interrupt handler', while Netscape described Windows as "a set of badly debugged device drivers". A description of Windows (any version), varies from "operating system" to a "thing that runs on top of a thing". Most modern operating systems are 'things on top of things', in that the core and shell are separate, and programs can run in the core (eg OS/2 CMD.EXE) or in the shell (PMSHELL.EXE).
IBM was pretty much the end of an anti-trust settlement, in the process, it had to out-source most of the IBM-PC to third-party providers. So it's not a great deal to implement the hard disk, etc. But IBM kept the BIOS for itself, and to clone the even then successful PC, one would need to clone the BIOS. I mean, you did not have to clone it. Tandy made things that were sort-of-IBM but not compatable: one could buy software for the Tandy 1000 or for the PC, as much as today, games come for the XBOX or PS-2. But to run stuff for the IBM Computer on a machine with a cloned BIOS was the ultimate quest, because all the neat stuff was coming out for the IBM PC. --Wendy.krieger (talk) 08:16, 23 October 2008 (UTC)

Article verbosity and general

This article is riddled with grammatical errors, with each sentence constructed in a seemingly awkward fashion. It is actually quite difficult to comprehend what the author is attempting to convey. I would recommend a complete re-write, but I am too lazy to do so myself. Additionally, I will most likely not return to this page to check on any responses to this comment, so don't bother responding to tell my to "do it myself" or the like. Thank you. —Preceding unsigned comment added by 24.4.236.15 (talk) 05:33, 10 May 2008 (UTC)

I agree, the English is horrible. In addition, it's written more like a lecture than an encyclopedia article. I'm not an expert on the subject matter or I'd rewrite substantial parts of it myself. 66.60.236.119 (talk) 09:17, 22 November 2008 (UTC)
Wow. I agree. "Many people do not realise that..."? This page definitely needs a rewrite. 74.183.191.55 (talk) 05:07, 15 December 2008 (UTC)

MS-DOS 6.0 source code

I think that someone can add information about fact, that MS-DOS 6 source code is available to download. What do you think 'bout that?? 83.23.243.25 (talk) 23:45, 26 December 2008 (UTC)

It isn't that I know of. Do you have a source of that information? JeremyMcCracken (talk) (contribs) 00:29, 27 December 2008 (UTC)
Of course, that I have a source.

Here it is: http://digg.com/security/MS_DOS_6_0_appears_to_be_on_Google_Code_Search And here you can download MS-DOS source code: http://www.nurs.or.jp.nyud.net:8080/%7Ennn/MS-DOS.6.0.Source.Code.zip

Info about size:

Compressed source Uncompressed source
~20 MB ~60 MB

That's all. I hope that it will be added soon. Mkmk101 (talk) 21:11, 27 December 2008 (UTC)

P.S. : Merry Christmas and a Happy New Year!

Unfortunately it's already been pulled from both of those links; I guess the guys at MS caught it about the time google did. Someone on the digg page was raising questions about whether it was real or not; without being able to get to it we can't verify that it was real. (I notice someone says MS-DOS was done entirely in ASM on Xenix, while this code is a combination of ASM and C. OpenDOS is mixed in that way; I wonder if someone took it and replaced the instances of "OpenDOS" with "MS-DOS".) JeremyMcCracken (talk) (contribs) 22:16, 27 December 2008 (UTC)
The second link still works fine for me. Ghettoblaster (talk) 23:02, 27 December 2008 (UTC)

vista dos

Need more information on this please --Something1235

There is no DOS in Vista. Command Prompt (Windows) SchmuckyTheCat (talk)
All versions of Windows NT (including 2K, XP, 2K3, Vista, 7), in the x86 form, continue to support the DOS emulation. Commands such as debug, edlin, edit, are DOS commands, and will run natively under MS-DOS 5.0. Edit sis the version that comes with Windows 95 and later, while debug comes from MS-DOS without modification. --Wendy.krieger (talk) 06:15, 11 April 2009 (UTC)

Hard drive size and capacity limits for DOS (why is this not mentioned here?)

Something very basic is missing from this wiki article on DOS.

DOS stands for Disk Operating System, yet no mention is made regarding DOS (or ms-DOS) and it's ability (or not) to be able to correctly access (read and write) hard drive volumes up to the maximum capacity of the FAT32 specification.

No mention of DOS and it's relationship to the BIOS int13 and enhanced int-13 function calls.

No mention if DOS can reliably read and write files to large hard drives (exceeding 128 gb) or that are formatted with non-standard cluster size. For example, if I have a 500 gb drive formatted as FAT32 with a 4kb cluster size, I'll have 120 million clusters (well beyond the so-called maximum as stated by Microsoft). What is known about ms-DOS and it's functionality with such a drive? —Preceding unsigned comment added by 74.12.203.91 (talk) 02:55, 17 April 2009 (UTC)

MS-DOS uses pure int13 to talk to IDE drives via CHS. This might be an unaddressed topic for this article. Addressing large IDE (not SCSI) drives that do not match CHS via int13 is a BIOS issue, not MS-DOS, and is extremely technical and not an encyclopedic topic.
MS-DOS died with FAT16 which had a maximum disk size of 2GB, and problems with CHS were present at much smaller sizes than 2GB. MS-DOS uses pure INT13 to talk to a drive. Int13 repeatedly came up against addressing limits as drive sizes increased and would need some sort of Logical Block Addressing to see the whole drive, either through the BIOS or software that loaded before DOS and translated the disk.
The real mode bootstrapper in Windows 95 OSR2 and later, which could read and write FAT32, had the same issues requiring LBA.
LBA issues aren't unique to MS-DOS, but other operating systems don't use real mode INT13. Win9x, WinNT, Linux, and any other OS all have LBA handling in their kernels, and strange things happen with partitions larger than certain sizes on certain kernels.
Formatting a modern size drive as FAT32 with 4k clusters is outside the specification for FAT32. It's broken. It's as broken as giving MS-DOS a 4GB drive formatted FAT16 with 64k clusters. Broken things aren't encyclopedia topics. Also, MS-DOS is dead and wondering how it works on any modern hardware is not an encyclopedic topic. SchmuckyTheCat (talk)
There is some about DOS and disk access at DOS; the limitations of both FAT16 and INT13 apply to most DOS-type systems that were manufactured. JeremyMcCracken (talk) (contribs) 16:48, 19 April 2009 (UTC)

formatting disks

When formatting CD's in XP MSDos and it comes up to ask what is the volume label, what is that? Please advise. I'm not that familiar with Dos. Thank you. pthreads10 Pthreads10 (talk) 21:09, 22 May 2009 (UTC)

See volume label; think of it like the name under the disk icon in "my computer". JeremyMcCracken (talk) (contribs) 06:08, 24 May 2009 (UTC)

Merge from MS-DOS 6.22

I didn't place the tags, but I support merging. There isn't much content in that article, and I think it belongs best with the discussions of versions here. The tag is in the "End of" section, but I think it would be best at the end of the section above. There's a little discussion of the 6.x changes in the "Legal issues" section; maybe put some of that up after the MS-DOS 5 discussion. JeremyMcCracken (talk) (contribs) 22:54, 19 June 2009 (UTC)

MS-DOS 7.0, 7.1 and 8.0

Do these operating systems ever refer to themselves by these names? I have Windows Me and when I type VER at it's DOS prompt it says "Windows Millennium [Version 4.90.3000]"

There is a DOS function (not a shell command but a function to be used inside .EXEs) that returns a DOS version number to a program (the exact function number eludes me ATM). That's where these numbers come from. -- 85.179.170.121 13:28, 12 May 2007 (UTC)
True. Programs would check that int 21 call to see what version was running, and it's maintained for that reason. While the products aren't actually released as "MS-DOS", the code does say "I'm DOS v 7". JenKilmer 16:46, 29 May 2007 (UTC)
If you run something like MSD.EXE (from DOS 6.22) or Manifest (QEMM), or even 4DOS, the true DOS version is displayed. This is the one set in the DOS kernel, not just a string that command.com prints. DOS in Windows 95a is 7.0, 95c, 98, 98SE is 7.1, and WinME is 8.0. One notes that both IBM-DOS 6.1 and PC-DOS 6.1 return 6.0 in this environment. Wendy.krieger 07:27, 30 August 2007 (UTC)

Actually MS-DOS version 7.X and MS-DOS 8.X were also by them selves but these versions were not in the open market. These were integrated in Windows 9x. In fact I'm actually running it in a Virtual PC 2007.

So how can I find out the true MS-DOS version of my current OS, Windows XP Pro; SP3 --John R. Sellers (talk) 21:30, 16 December 2008 (UTC)
You can find the true DOS version, by running a program that reports it. Windows NT runs DOS 5.00 (edlin,exe and debug.exe from NT (eg XP), are bytewise equal to that in MS-DOS 5.00a (ie 1991.11.11 date). --Wendy.krieger (talk) 08:11, 27 June 2009 (UTC)
The DOS version would be 8, since it's newer than Win ME. JeremyMcCracken (talk) (contribs) 22:12, 16 December 2008 (UTC)
I assume that would be the major version number, and the minor version number would be something like "1c", making it Microsoft OS 8.1c? So Windows 7 would be Microsoft OS 10, right? --John R. Sellers (talk) 21:18, 14 January 2009 (UTC)
There isn't anything other than the eight, because the DOS version is no longer tied to the Windows version. It was incremented while DOS-based windows versions were produced, but it stopped with DOS 8 with Me. After Me, it was only a compatibility object for the DOS subsystem (NTVDM) and placed on the boot disks, so it's never been upgraded. JeremyMcCracken (talk) (contribs) 22:25, 14 January 2009 (UTC)

MS-DOS 7.11 was inside later version of Windows 95 and in Windows 98 and Windows 98 SE. MS-DOS 7.11 had USB support which was the reason why some version Windows 95, Windows 98, and Windows 98 Se supported these buses. —Preceding unsigned comment added by 71.109.71.106 (talk) 20:47, 22 November 2007 (UTC)

There never was an "MS-DOS 7.11". It's 7.10 in all three versions (OSR2, 98FE, 98SE). Native DOS does not support either USB or Long file names. Apart from a few files like IO.SYS, DOS was not upgraded through the three versions.--Wendy.krieger (talk) 07:39, 10 July 2009 (UTC)

Citation for Unreferenced discussion of MS-DOS 7

It states in the topic about MS-DOS 7 once being created as a stand-alone project, although no citation available. It's been since April 2008 since a ref? was created. The only site I could find any information about it was on Computer Hope. It states "The latest stand alone version of MS-DOS is 6.22. Microsoft Windows 95 and 98 included a version of MS-DOS known as 7.0. While a user is capable of utilizing MS-DOS 7.0 through Microsoft Windows, MS-DOS 7.0 cannot be purchased as a stand alone version and requires Microsoft Windows be installed to run."[1] However, IBM did release a version 7 of their PC-DOS in 1995, possibly to rival MS-DOS 7.0. But once again, these are assumptions, which have been in this article for about a year now.

tldr; Can we only assume that Microsoft was going to create a stand-alone copy of MS-DOS 7, rather than make assumptions and stick it in the article? I believe we should remove this information altogether by a certain date, and state that MS-DOS 6.22 was the last stand-alone copy Microsoft released, and they removed it from marketing on November 30, 2001. Troyoda1990 (talk) 23:08, 8 July 2009 (UTC)

A couple things. First, there was something that was brought up somewhere on a talk page, but I can't find it now, about an MS-DOS 7 that someone else made from Win 9x with the BOOTGUI option off. It's called "MS-DOS 7.1" or "MS-DOS 7.10"; google "MS-DOS 7.10" and you'll find some forum discussions about it, and this link appears to be the project home page. Separately is this link, which is a document from MS that was produced in the antitrust case; it would seem to support that a MS-DOS 7 from Microsoft was at least discussed. JeremyMcCracken (talk) (contribs) 02:13, 9 July 2009 (UTC)
There was never any serious discussion to make a standalone MS-DOS 7. That last document discusses ways of taking the Chicago kernel and architecture without a GUI - a Win32 console. It also doesn't take any of those idea seriously it was unfeasible and uninteresting and wouldn't gain any meaningful market. SchmuckyTheCat (talk)
Microsoft did prepare two betas of "MS-DOS 7.0", based on 6.00 and 6.22 code. The former is still found around the internet, while Schuhmann "Undocumented DOS" refers to a MS-DOS 7 with Drivespace in it. Microsoft was playing a wait and see, to see if there was enough interest in the already released DR-DOS 7 and PC-DOS 7, to see if the DOS market would survive Windows 95 release. However, we note at this time, magazines like DOS Tools went out of publication, and much of the world took up Windows 95. MS-DOS 7.0 was never released as a stand alone product.--Wendy.krieger (talk) 07:35, 10 July 2009 (UTC)
MS-DOS, as the boot component of Windows 95, has Drivespace. MSDOS 7 as a beta product never existed outside of Windows 95. This stuff is not important enough for Wikipedia, particularly the stuff based on crap found on the Internet and not documentation. SchmuckyTheCat (talk)
I'm pretty sure a memo from Microsoft produced during the MS antitrust case would meet WP:RS. JeremyMcCracken (talk) (contribs) 22:45, 10 July 2009 (UTC)
The only documentation I could find referencing MS-DOS 7.0 is on Microsoft, only for an Error Message on Windows 95. Also, over-reading the United States v. Microsoft Anti-Trust Case transcript, its only reference to MS-DOS states that a function was not available for printers like in Windows, which has nothing to do with marketing or product placement of MS-DOS Version 7.0 as a standalone product. Thus, we should only remove the sentence until a reliable source is found. This is like saying Microsoft originally planned to create another 9x-based operating system after Windows ME as a standalone product, but decided to focus only on Windows XP for home users. Troyoda1990 (talk) 18:08, 11 July 2009 (UTC)
I meant the link in my first reply here; seems to be a discussion over a possible version 7 that eventually became improved DOS compatibility in Win95. JeremyMcCracken (talk) (contribs) 19:40, 11 July 2009 (UTC)
Oh, well, I tried clicking on that PDF File, along with this one being the same, and it gave me a 403 Forbidden page. However, if you do believe that the pdf file can backup the statement of "MS-DOS 7 was considered for shipment as a standalone product," then by all means, we should use the PDF as a reference for the statement. Troyoda1990 (talk) 04:24, 12 July 2009 (UTC)
There was never any serious discussion to make a standalone MS-DOS 7. That document doesn't take any of those ideas seriously; it was unfeasible and uninteresting and wouldn't gain any meaningful market. There is no indication this is anything other than a low-level project manager brainstorm. It is a huge leap without contextual justification to go from this document from a small group or single PM to engaging an entire corporate division to "consider" a new product. SchmuckyTheCat (talk)
OUTDENT Not sure why the PDF gave you trouble, site hiccup maybe? I'd link the google HTML version, but it's pretty mangled. As for this as a citation, "unfeasible and uninteresting and wouldn't gain any meaningful market" isn't anything citable here. It may be true, but it's OR. Of course, keep the context of this memo in mind- it's dated Feb 1993, when Win95 was two years from release (indeed, it's referred to as Chicago). The concern is whether it should cite "although at first DOS 7 (which was the DOS part included in Windows 95) had been developed as a standalone OS", which it doesn't, because that says "developed". I'd recommend changing the first paragraph to read
MS-DOS has effectively ceased to exist as a platform for desktop computing. After initial discussions of a possible MS-DOS 7,(cite) it eventually became fully integrated into Windows 95, used mostly for bootstrapping, and would never again be officially released as a standalone DOS.
I think the memo is a sufficient citation for that. JeremyMcCracken (talk) (contribs) 20:52, 12 July 2009 (UTC)
Here is another link for that PDF: [3] Some discussion on it would be grand before changing this section. JeremyMcCracken (talk) (contribs) 21:02, 12 July 2009 (UTC)
The second line worked this time, for some reason I keep getting the error. Anyways, I would like to second the paragraph suggestion. Since you came up with it, you should add it yourself to get credit for it, but as always, it's best to leave the topic open for discussion. Troyoda1990 (talk) 03:42, 13 July 2009 (UTC)
I also reread through it and I found it interesting how they called some consumers "Windows Haters." It's like they separated Windows, MS-DOS, and Chicago into three divisions, while Windows was still based off the [Program Manager] interface. They also had multiple plans for how to go through with MS-DOS 7. Very interesting read. Troyoda1990 (talk) 03:47, 13 July 2009 (UTC)
I oppose using this as a reference. One single "idea" document from a small group is not evidence that Microsoft (meaning the company as a whole) had any interest in a standalone DOS 7 product. If the document had come from a major executive and contained some directives to explore costs for doing any of these options, that would be different. SchmuckyTheCat (talk)
Then should we just remove the sentence altogether, or reword it to make sense to the time era of the release of Windows 95? For example:
MS-DOS 7 was developed along with Windows 95, the successor to Windows 3.1, and as a platform for desktop computing, Windows 95 was much easier to use Microsoft's command-driven operating system. As a result, MS-DOS 7 was integrated into Windows 95 and its kernel, combining the separate products into one.
Yes, it's worded weird, but I'm getting the overall idea out. The statement states nothing about development of MS-DOS 7 as a standalone product, but the phrase I wrote about 95 being much easier to use makes sense to why it wouldn't be. Do we really need to say that MS-DOS 7 was considered to be a standalone product? It's like saying Windows XP SP2 was such a great update, it was considered to be sold as a standalone product like Windows 98 SE. Those are assumptions, which is what the sentence written is. We can just state that MS-DOS 6.22 was the last copy to be sold in stores. Troyoda1990 (talk) 17:15, 14 July 2009 (UTC)
I searched the history and found an IP had added that statement about a year ago, and it was immediately tagged with cn. I just went through the article doing some copyediting and moving and I just reverted back to the text from a year ago, then re-wrote that.
Other changes I made were to make things more clear, more flow along a timeline, remove references to "recent" and "latest" when referring to versions of Windows released eight years ago, etc. SchmuckyTheCat (talk)
I was on the fence about whether it was worth mention or not, so might as well go without it. JeremyMcCracken (talk) (contribs) 04:31, 15 July 2009 (UTC)
Looks good! I'm glad we covered this section. I never knew one line would cause so much drama on Wikipedia's Talk page. haha. Troyoda1990 (talk) 03:56, 17 July 2009 (UTC)

Windows NT/2000/XP COMMAND.COM Version (Bug)

This is not really a bug. In WinXP (and presumably 2000, 2003, Vista, etc.) there is a VER.EXE file in C:\WINDOWS\SYSTEM32 which displays the Windows XP version number 5.1. COMMAND.COM finds a file VER in the search path, and executes it, displaying "Microsoft Windows XP [Version 5.1 . . . ]". For some reason, when COMMAND is called with /c or /k it executes the internal ver command (which is normal DOS behavior); but when VER is typed it searches for VER executable files, instead of using the internal command (weird).

Command.com is the same in Win2k and WinXP. There is no VER.EXE, either. What happens is that when you run a DOS program, commands are passed through to the underlying CMD.EXE or 4NT.EXE. You can see this by running dir /2 (an extended 4nt switch) under both of these.
When command.com is run with /c or /k, the environment created is to allow programs to communicate with assorted TSR programs, and here it needs a DOS-like string to answer to the programs. This is why it returns strings like MS-DOS 5.00.500. Wendy.krieger 07:45, 30 August 2007 (UTC)

Correct me if I'm wrong, but the version 5.0 or version 5.1 refers to the Windows NT version, not the DOS version. I seriously doubt that Windows NT contains the DOS 5.0 or 5.1 command.com. 71.159.225.121 (talk) 23:14, 3 March 2008 (UTC)

The version of DOS in Windows NT (all versions), is "5.00.500". Unlike OS/2 (where DOS version is OS/2 version is 20.xx) it has nothing to do with the underlying NT. The dos from Windows 2000 is bitwise equal to the one in windows XP, although vista has a different ntbio.sys and command.com. --Wendy.krieger (talk) 07:15, 4 March 2008 (UTC)

I suspect that the version of Dos in WinXp, etc, is roughly the same as in Windows 98 or Windows ME. The part about it being DOS 5 is certainly incorrect, although which version actually version it is hasn't been identified. For instance, you can type mem /c, which was a feature first created in dos 6.0. I rewrote the section to remove the incorrect claim. Xeos (talk) 03:19, 21 February 2009 (UTC)

No, the command.com in all versions of Windows NT forked from DOS 5. Some "new" commands or features were backported, but certainly not the stuff in win9x. SchmuckyTheCat (talk)

Calling it dos 5 is still clearly incorrect, and beside the point. Even if the code is based on DOS 5 (as is, for instance, dos 6), it's been heavily changed. If there was a fork (and I would like to see an external citation as independent evidence of this), that itself might be worth noting on the article, but calling it Dos 5 is grossly misleading. Xeos (talk) 16:17, 27 February 2009 (UTC)

I wonder if command.com simply hasn't been changed since 5, just updates and changes to the accompanying utilities. JeremyMcCracken (talk) (contribs) 16:19, 27 February 2009 (UTC)

Apparently my memory is incorrect - dos 5 did have the mem command. But according to this page, http://www.textfiles.com/hacking/MICROSOFT/dosundoc.txt dos 5 had a ver /r switch, and the so-called dos 5 in WinXP, et al, does not. If the dosundoc.txt file is to be believed, there is actually lots of evidence that the DOS in WinXp is not 5, or 6, even. Try dir, which is supposed to show hidden files in 5 and 6, but does not in command.com under Winxp. Xeos (talk) 16:57, 27 February 2009 (UTC) Well, egg on my face, if you call ver /r or dir, using the command.com /k switch, you get the proper dos 5 or 6 behavior. just starting command.com gives you strange intermediate results. Xeos (talk) 17:02, 27 February 2009 (UTC)

I have a PC dos 5.0 disk and it's command.com file is 47,987 bytes. The windows command.com file i s50,620 bytes. But, PC dos isn't necessarily MS-dos. Xeos (talk) 17:19, 27 February 2009 (UTC)

The thing about "DOS version x.xx" is that there are particular system calls (functions 30h and 33.06h) which any program can look at the DOS running. WinNT DOS reports 5.00 and 5.50 here. Programs that require DOS 6.0 or greater won't run, while all of the utilities of PC-DOS or MS-DOS 5.00 will happily run here. I have a desktop icon pointing to PC-DOS 5.00 command.com, without modifications, that runs quite happily in Windows NT. Edlin.exe and debug.exe are indeed identical in MS-DOS 5.0 and in Windows 2000.
The internal DOS version does not necessarily match either what Command.com or the box says. We have, eg a product IBM-DOS 6.00 with command.com saying 6.10, and truever 6.00. PC-DOS 6.10 has command saying 6.10, truever 6.00, MS-DOS 6.21 says command.com 6.21, ver 6.20. and then of course the Windows 9x versions. --Wendy.krieger (talk) 08:08, 27 June 2009 (UTC)
It should be noted that the DOS version is really important to programs that expect to find DOS there. Any DOS 5.0 utility, either from PC-DOS or MS-DOS, runs without modification under any Windows NT/2K/XP/Whatever system. This is because the ntio.sys and ntdos.sys files call themselves MS-DOS 5.00 (or by a different call, 5.50). Programs that expect to find any different DOS, like 5.02, 6.0, 6.20, 6.22, 6.3, 7.0, 7.1, 8.0, will report an incorrect DOS version. So the DOS really is DOS 5.
You can't take the DOS debug utility (which expects to find DOS 5.0, since that's where it comes from) and use it on the DOS diskette produced by the diskcopy.dll utility (which produces a DOS that calls itself 8.0. You get incorrect DOS version. Edlin from NT/2K... will run on any version of MS-DOS, or any version 5.0, but not, eg PC-DOS 2000. --Wendy.krieger (talk) 11:26, 29 March 2010 (UTC)

Proposed merge-in of DOS

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
  1. ^ "Information about MS-DOS 7.0". Computer Hope. Computer Hope. Retrieved 8 July 2009.