|This is the talk page for discussing improvements to the MS-DOS article.
This is not a forum for general discussion of the article's subject.
|MS-DOS has been listed as a level-4 vital article in Technology. If you can improve it, please do. This article has been rated as B-Class.|
|WikiProject Microsoft / .NET / Windows||(Rated B-class, Top-importance)|
|WikiProject Computing / Software||(Rated B-class, Top-importance)|
|Threads older than 100 days may be archived by.|
- 1 DIR
- 2 Improve references in article
- 3 MS-DOS and Clones
- 4 changes
- 5 MS-DOS and Windows NT
- 6 Competition Section, Incomplete Thought
- 7 MS-DOS Version 2.2x and 3.35
- 8 Conflicting Release Date
- 9 Windows 95 up until Windows NT
- 10 Multi-tasking and "the other DOS 4.0"
- 11 Opening of source code
- 12 Localization
- 13 MS-DOS logo
- 14 Proposed merge with MIDAS (operating system)
- 15 Add stuff about XP.
- 16 Version history regarding 7.0 through 8.0
- 17 RAS syndrome
- 18 History
The "DIR" example needs to be changed. The output looks more like UNIX/Linux. I've never seen any such output from a DOS "DIR".
- Thats probably from a Windows machine running an updated version of DOS. Older DOSes, as you most rightly probably see, did not comma-separate nor tab out the display as nicely as the newer DOSes did. Dysprosia 12:22 26 Jul 2003 (UTC)
- It looked very strange to me. I replaced it with a heavily edited version of my root directory, from Windows 98. I stripped off the LFNs, I think it's pretty much the same as DOS now. -- Tim Starling 12:35 26 Jul 2003 (UTC)
- Well, comma separation came round v6.2, but I don't think we really need any of this to dig that far back :) Nice work Dysprosia
Improve references in article
|This article needs additional citations for verification. (February 2010) (Learn how and when to remove this template message)|
MS-DOS and Clones
MS-DOS is bought from a system that apparently is the root for DR-DOS as well, so it is unfair to talk as DR-DOS as a clone of MS-DOS. Because Microsoft were allowed to licence what they wrote for IBM for other manufactures, strictly speaking, MS-DOS is a clone of PC-DOS to run on computers that cloned the IBM PC.
The first retail or upgrade DOS was DR-DOS 5.0. All DOS versions before this were available as OEM only.
It's only after version 5, and the split between IBM and Microsoft, that people started to think of MS-DOS as the thing being cloned. In part it's more to do with the magic that Microsoft was putting into DOS (which they did not have a monopoly), and Windows (which they did), that other DOS manufacturers had the inside edge. IBM had access to the real DOS/Windows code up to DOS 6.20 (which is the basis of 6.30), and Windows 3.10 (which is the basis of Win-OS/2 3.1).
IBM were allowed to go it head on with Microsoft, so beginning with PC-DOS 5.02, IBM's DOS competed with Microsoft's DOS on the open market. When this arrangement finally lapsed, Microsoft released 6.22, and 3.11, where references to OS/2 as the server were replaced with Windows NT. Note that both OS/2 2.x and Windows NT rely on DOS 5.0 for their DOS utilities.
MS-DOS 6.x OEM disks are labeled "MS-DOS and additional tools". This split separates things like qbasic, memmaker, msav, msbackup, scandisk, as 'additional tools'. One could either licence DOS by itself, or with the additional tools. IBM licenced DOS itself, and provided its own editor, script language, antivirus, compression, etc.
MS-DOS 7, 7.1 and 8 exist, but as subsystems of the Win9x systems. The bootable DOS diskette produced by diskcopy.dll produces a DOS 8 diskette, however, the DOS utilities provided in Windows (eg debug.exe, edlin, command.com), are DOS 5. debug.exe is a straight byte for byte copy of the DOS 5 utility. Edit is from Windows 9x, runs under any DOS from 3.1 onwards. Likewise, the QBASIC from any Win9x cdrom is from MS-DOS 6.22, amd from Windows NT 3.x, 4.x, is straight out of 5.00.
I was expecting that there is a little changelog what changed in the different dos versions. Why is there nothing(only release date and versionnumbers) what features the different versions add/removed? mabdul 10:53, 16 April 2010 (UTC)
- DOS 5 (PC-DOS and MS-DOS), are essentially equal, except PC-DOS is compiled for the 386, while MS-DOS is a 286 version. This supports a new editor (QBASIC EDIT), a DOSSHELL based on Windows 3.0 standard mode, and an upper memory management program. The system is much more stable to 4.0 or 3.0. 2880 kiB floppies are supported.
- DOS 5 is the basis of the DOS emulation in both OS/2 2.x (except PPC), and NT i386.
- DOS 6 and Windows 3.x1 files are packed with compress v1, and with PKLITE. The install disks are now usable as program files.
- MS-DOS 6 seems to vary little, once one takes account of the version check. The differences between 6.00 and 6.22 and intermediate versions, amounts to some changes in IO.SYS, COMMAND.COM, and one or two other files. Between 6.0 and 6.2, there are changes in the redistributables (himem.sys, emm386.exe, mscdex.exe). The bulk of the differences between 6.20 and 6.22 are in the version check, although SETVER's table changes, and IO.SYS, COMMAND.COM are different.
- This version comes with a revised QBASIC (broken support for reading HELP.HLP), along with utilities similar to contempory DR-DOS 6.0. Memmaker (memory management), DRVSPACE (broken in 6.0, fixed in 6.2, removed in 6.21), DBLSPACE (new in 6.22), MSBACKUP, a broken antivirus MSAV.
- PC-DOS 6.0, 6.1 and 6.3 are very similar at byte level, and with the exclusion of version-check, more of the files become different. I did not decode the files to see if there are more files are equal save the version check. PCMCIA support is added.
- IBM did not licence Microsoft's 'additional tools', and included a back-version of PC TOOLS.
- MS-DOS 7 et sec are essentially identical, handle long file names in fat (when a driver is loaded). The bulk of the files are identical, save for IO.SYS, COMMAND.COM, and some files changing due to the introduction of FAT32. MS-DOS 8 is merely 7.2 rebadged. Very little changes here except what is needed for Windows (ie FAT32, loading straigt into Windows). Files packed with Extract into cabs.
- Most of the additional tools were removed (first to 'OLDMSDOS' on the cdrom), a new editor replaced the QBASIC editor.
- PC-DOS 7 is basically PC-DOS 6.3, with an updated editor and rexx support - these appeared first in the DOS 6.3 emulated in OS/2 PPC. As with DR-DOS, IBM includes Stacker 4 (not the DOS and OS/2 version). In line with SAA, the files are packed as for OS/2, using the PACK2 utility. PC-DOS 2000 is a cosmetic change to PC-DOS 7 + bugfixes.
- PC-DOS 7.1 is a version for building installs, supports FAT32, but not LFN, fits essentially onto a floppy disk.
MS-DOS and Windows NT
Somewhere, it should be noted that people call the Windows NT command prompt 'DOS'.
Microsoft did a lot to foster this. The icon in Windows NT 3 and 4 for cmd.exe is the same as the MS-DOS 5 icon, and the same window can run both DOS and Win32 command-line utilities. Given that Windows cmd.exe is more like command.com than OS/2's version, and that its command.com actually passes commands through to cmd.exe (or any other underlying command processor) to process, gives the illusion that Microsoft's command window really is DOS. In Consumer windows (win9x), it actually is DOS that processes it, but this command.com pretends to be Windows.
Contrast this with IBM's OS/2, where this confusion never arose. One has specific DOS and OS/2 command-line windows, each with their own set of programs. One does not read in the OS/2 press (eg OS/2 magazine), any confusion about the OS/2 command-line being 'DOS'.
It's not that people would confuse the two - this did not happen in OS/2, but that Microsoft made this easy to happen.
- People even today refer to the CLI as DOS. ( CompTIA A+ 20-701 ). The article propagates this also: "MS-DOS is still alive today in Windows 8, but in the form of Command Prompt." I thought that in Windows NT 4, that command.com generated a stack error, but I can test it with Windows 2000, NT v5.
I would rewrite this line: "MS-DOS is still alive today in Windows 8, but in the form of Command Prompt." as "The Command processor of DOS, has evolved to be compatible with the latest file systems, and Operating systems." ( You can do a dir with OS 8 on a ExFAT ). I would fit the word version in there too somehow... — Preceding unsigned comment added by 220.127.116.11 (talk) 10:51, 6 February 2013 (UTC)
This is a huge mistake. People still call a console window a DOS box and similar terms. The two are totally different; the programs that execute in the two are different. This article should not promote the confusion that a console window executes DOS programs. Sam Tomato (talk) 16:31, 6 May 2015 (UTC)
It is not true that "the same window can run both DOS and Win32 command-line utilities". A DOS box can only run DOS applications and a Windows console window can only run Windows console programs. There are versions of DOS commands (such as "xcopy" that first appeared in DOS) available for console windows but they are not the same program. Sam Tomato (talk) 17:58, 6 May 2015 (UTC)
Competition Section, Incomplete Thought
In the Competition section is the sentence:
I am having trouble seeing what completed thought is in the phrase after the semicolon. What thought, from the first part of the section (before the semicolon), is being commented upon / added to / whatever in the last part (after the semicolon)? The last part needs to be reworked to clarify what it is saying. — al-Shimoni (talk) 23:39, 13 December 2011 (UTC)
- I rephrased this a bit. Hopefully it becomes clearer now. --Matthiaspaul (talk) 23:58, 13 December 2011 (UTC)
MS-DOS Version 2.2x and 3.35
Hello. My name is Ole Juul and I live in Coalmont BC. I am an avid user of MS-DOS and an amateur computer historian. I would like to point out that the Wikipedia list of MS-DOS versions includes "Version 3.35 (OEM)" but that there does not appear to be any proof of that version's existence. To be included in this list I strongly believe that there should be some official Microsoft documentation of (at least) intent, or an actual copy of the software. As it stands, it would appear that there is neither.
The only mention that I can find is on a Russian "abandonware" site. See here: http://oscollect.old-dos.ru/OS/MS-DOS/Final%20Releases/
Unfortunately that version is actually 3.3 and _not_ 3.35 like it says. Another "abandonware" source came from Hungary and it was also just a mislabled version of 3.30. There are legitimate sellers of used and historical software as well but I have not been able to locate one with any mention of version 3.35. It would appear that the problem could be a typographical error somewhere along the way, and that it got carried forward. There are now other lists on the web which reflect this apparent error. Is it possible that Wikipedia has taken the information from another web site without any verification?
I have checked with others in the DOS community, and the version is not known. We all collect any historical versions of DOS and it would appear that a copy of 3.35 has not surfaced. If someone here wishes to confirm a version of 3.35 you can view the code in the COMMAND.COM file and the version number will be visible about four screens down (~line 99). The Russian version says 3.3. You can also type "ver" to get the version information - but you will obviously have to boot with that particular version in order to do that.
I hope this information will be helpful in sorting out this situation. Having Wikipedia list something which does not appear to exist is causing some people to spend a lot of time on a wild goose chase. I would challenge the original poster of the list to actually produce a copy, because I do not believe that this list has any legitimacy without the compiler either referring to Microsoft documents, or looking at a real world copy of each version. Otherwise it is hearsay and, unless labelled as such, should not be included.
Ole Juul oj (at) coalmont (dot) net
PS: I am unable to register because my name, and anything close, is already registered. My sense of honesty will not allow me to use a pseudonym. However, I have great trust in the integrity of the editors here and have every confidence that this matter will be dealt with appropriately. 18.104.22.168 (talk) 11:25, 6 March 2012 (UTC) Ole Juul
- Really? Typing User:Ole Juul in the search box gets you to a "user not registered" page. WP:Be Bold and change it. Trusting in the integrity of editors here is going to get you disappointed. --Wtshymanski (talk) 14:29, 6 March 2012 (UTC)
- Right now, I have no personal recollection of a MS-DOS or PC DOS 3.35, but haven't checked my archive yet. It is possible that it existed, but if it did, it's rare and definitely not one of the "mainstream" releases. There was alot of parallel development and quite some confusion in version numbers in the time frame between 3.3 and 5.0, and OEMs sometimes used their own numbering schemes.
- The version number must not necessarily be reflected in strings embedded into COMMAND.COM (although it does so most of the time). Have you checked the other system files, INSTALL/SETUP or the likes or README files as well? Have you compared the files in the mentioned "distribution" with files from a normal MS-DOS 3.3 issue? Are there any differences? Have you checked the version number reported by the system on INT 21h/AH=30h level? (IIRC, MS-DOS 6.21 continued to display as 6.2. And PC DOS 6.1 reported a version number of 6.0 on API level. Strange things happen... ;-)
- FWIW, there was for sure an OEM version of DR DOS 3.35 in October 1988, but DR DOS is not based on MS-DOS, of course.
- Regarding your registering problems, once you have registered you could set up a signature different from your account name. --Matthiaspaul (talk) 15:01, 6 March 2012 (UTC)
- It was determined in a Vintage-Computer.com forum that the 3.35 version Ole had was actually just a DOS 3.3 that was renamed.
- I found a great resource with some DOS history, but am unsure how best to integrate it into the wikipedia article -- maybe someone else can run with it? The link is http://www.os2museum.com/wp/?page_id=639 --Trixter (talk) 02:35, 18 March 2012 (UTC)
There was never a MS-DOS 3.35 - removed from version list. I'm 99.9% sure there was never a MS-DOS 2.2x either, any evidence to the contrary ? If not then it should be removed as well. Asmpgmr (talk) 22:43, 12 June 2012 (UTC)
- Confirmed, I was a computer historian too, and I never saw a MS-DOS 3.35. 3.31 came with the IBM Portable. As for MS-DOS, PC-DOS was version 2.0 and 2.1, MS-DOS was 2.0.1 and 2.1.1. PC-DOS 2.2 was a patched version by Tech-PC Magazine that fixed a register order error. ( you could rename PC-DOS to be 2.1 ( no change ) or 2.2 or 2.1p ) Ill look for others. — Preceding unsigned comment added by 22.214.171.124 (talk) 10:55, 6 February 2013 (UTC)
- FWIW I vaguely recall a Japanese OEM DOS 2.25, it supported AVAILDEV and SWITCHAR in CONFIG.SYS. –Be..anyone (talk) 03:45, 1 April 2014 (UTC)
Conflicting Release Date
In the current article, it says on the side box that MS-DOS 1.0 was released in 1981 (this seems to be correct, as I'll point out later), but on the text itself it claims to have been released in 1982. I've checked the source of that, and can't find any solid evidence that it was in 1982, just some people talking about stuff happening "months later" or "a year later" with questionable context.
However, after checking the fourth External Link in this page (http://www.linfo.org/ms-dos.html) the article clearly states a more accurate date, August 1981.
So, I ask, is it acceptable to change the article to reflect this, and accept August 1981 as the proper Release Date?
- Please Please do so. I was there, and I am the reference, so I, by rule cannot source it. DOS 1.0 came out with the first PCs when they were shipped. It only supported single sided disks, and its file creation dates were Aug 04, 1981. Later on in the same month in August, My brother and I went down to the computer land in El Cerrito, CA and purchased a dual DSDD IBM PC with 256K of ram, and DOS 1.1, which was necessary to handle the second side. So August, 1981 saw two versions of DOS released.
- Lets see about solid evidence. The Disk labels for DOS 1.0 and 1.1 both say (C) 1981. DOS itself says (C) 1981, and ....
- But a picture....
- It Reads: "IBM Personal Computer Computer Language Series 6172212 DOS Version 1.00 (C) Copyright IBM Corp. 1981 Licensed Material Property of IBM
- -- 2013-02-12T02:26:16 126.96.36.199
- Things are complicated a bit by the fact, that there were two operating systems, MS-DOS and PC DOS. PC DOS 1.0, an OEM version of MS-DOS tailored specifically for the IBM PC, was released in August 1981 together with the IBM PC, but MS-DOS itself was not available to end-customers at this time (or at least I'm not aware of it and could not find any sources for it). PC DOS 1.1 was released in May 1982 together the double-side floppy disk upgrade of the IBM PC, and it was only in June 1982, when other OEM versions of MS-DOS started to appear, and while some of them maintained the same version numbering schemes, other OEMs used very different names and versions, sometimes causing alot of confusion when it comes to the task of determining the exact MS-DOS version they were originally derived from. --Matthiaspaul (talk) 18:10, 12 February 2013 (UTC)
Windows 95 up until Windows NT
The statment "Since the release of Windows 95 up until Windows NT, it was segregated as a full product used for bootstrapping, troubleshooting, and backwards-compatibility with old DOS games and no longer released as a standalone product." doesn't make sense. Windows NT was released in 1993, two years before 95. Emmette Hernandez Coleman (talk) 11:04, 2 November 2013 (UTC)
- I tore this section apart, clarified a little stuff (concurrent NT/9x), removed some stuff that was replicated elsewhere in the page. I don't think this bit is an issue anymore. --Overand (talk) 15:51, 19 June 2015 (UTC)
Multi-tasking and "the other DOS 4.0"
It might be interesting to add some info about the "multi-tasking DOS 4.0" described in this blog post. It seems that this was developed alongside 3.0, intended to be a major new version with advanced multi-tasking features. But it was later abandoned in favour of OS/2, except for cut-down releases for some OEMs who'd already signed up for it. What most people know as MS-DOS 4.0 was actually a different product, released by IBM as PC-DOS 4.0 some time later, and the multi-tasking features never made it into the main line of DOS. - IMSoP (talk) 00:43, 13 March 2014 (UTC)
There is Multiuser DOS for this purpose, if the article doesn't link to it anywhere please add it, e.g., put it in MS-DOS#See also.No, that's something else. We only have a shaky redirect from European MS-DOS to New Executable. –Be..anyone (talk) 17:05, 13 March 2014 (UTC)
- Hm, there's a few interesting references scattered about, and I think it deserves its own article, so I've gone ahead and stubbed one out: Multi-tasking MS-DOS 4.0.
- I'll go around linking it in and fixing up redirects now. - IMSoP (talk) 19:32, 15 March 2014 (UTC)
Opening of source code
Note that the source code is not even open source, as the policy is "see, don't touch", and applies to DOS 1.0/2.0, and under a restrictive research license (non-commercial, not for distribution and modification, etc). -Mardus (talk) 14:03, 27 March 2014 (UTC)
Stuff belonging to the localization apart from the referenced COUNTRY.SYS, all covered in the same TechNet MS DOS 6 technical reference: DISPLAY.SYS (handles codepage changes for the display), various *.CPI files including EGA.CPI (display fonts), KEYBOARD.SYS (keyboard layouts), KEYB.COM or EXE, NLSFUNC.EXE (manages CHCP for keyboard/display/printer), CHCP (internal command). Lots of codepages existed, and the NLSFUNC stuff allowed to use any of the "prepared" codepages with CHCP, not only two as in OS/2. –Be..anyone (talk) 03:04, 1 April 2014 (UTC)
- Here is a screenshot of MS-DOS 6.22 set to German settings:  As you can see the language is still English even after a reboot. Your source also says "enables MS-DOS to use country-specific conventions for displaying times, dates, and currency". It does not say anything about User Interface language. Additional screenshots are also available if you wish.
- May I remind you that the burden of the source is on you? But so far your have been reinstating your change without additional source, justification or anything; while it was I who took my time to take screenshots. Are you familiar with out edit warring policy?
I am hoping Sk-Eiht is viewing this, because I reverted one of his edits that I found objectionable, per WP:BRD. In view of the recent attempts to place an image in the logo area of the article, I thought it is time we had this discussion.
Logo is different from cover art, seal, splash screen and mascot, although computer icon is a form of logo. For example, File:Office logo (Pre-2003).png is a logo of Office 97. File:Microsoft Office 97 Professional Box Art 2.jpg is a cover art but not a logo. Another example is ESET NOD32: File:ESET antivir 7 logo.png is its logo but the software recurrently shows a picture of robot, as seen in File:ESET-NOD32 Antivirus.JPG. The robot is not the logo.
Does MS-DOS have a logo? If I remember correctly, MS-DOS didn't have a logo. It did have a cover art, but that was just art and was not registered as its mark. File:Microsoft DOS.svg is a definitely a logo and is derived from both Windows 3.1's DOS prompt logo and the cover art, but its source does not indicate that it is made by Microsoft.
P.S. Welcome to Wikipedia, Sk-Eiht!
- Sorry for my edit, it was absolutely in good faith. I thought it was the official logo of MS DOS. I am not a native English speaker (I am Italian), but I got what do you mean. Thank you for your kindness and nice work on wikipedia :) --Sk-Eiht (talk) 17:25, 4 September 2014 (UTC)
Proposed merge with MIDAS (operating system)
Add stuff about XP.
- Yes, but the kernel of Windows XP is Windows N.T. not M.S.-D.O.S. so the information is not relevant to this article, and the Command Prompt article itself has enough about it.
- Sincerely, --Namlong618 (talk) 11:07, 23 February 2015 (UTC)
Version history regarding 7.0 through 8.0
Microsoft never officially made any operating systems named "MS-DOS 7.0", "MS-DOS 7.1", nor "MS-DOS 8.0". Instead, Windows 95 included a version of DOS which reported itself to applications as version 7.0. Windows 95 OSR2.x (95B/95C) and both versions of Windows 98 included a version of DOS which reported itself to applications as 7.1, and Windows ME further still included a version that reported itself as 8.0. Reading an archived talk page, the information very much smells like original research and the version numbers themselves should really have a reliable source to cite from (preferably Microsoft themselves).
As there were no such stand-alone operating systems, MS-DOS at that point in history being included in Windows 95, Windows 98, and Windows ME as more of a compatibility object than the primary system itself (arguably--the software architecture probably made it prohibitively complex to fully decouple from DOS, but that's a moot point here). The newer version numbers along with the tie to the Windows 9x OSes can be noted with a proper citation, but I don't believe they should be represented as whole operating systems on their own right. --Chungy (talk) 15:01, 21 November 2014 (UTC)
"This article is about Microsoft's MS-DOS. For other compatible operating systems of the DOS family, see DOS." MS-DOS already means Microsoft DOS, so isn't the above equivalent to "Microsoft's Microsoft DOS"? 188.8.131.52 (talk) 20:27, 17 March 2015 (UTC)
- I don't know the TLA RAS, but your shorter DAB note is okay ;-) –Be..anyone (talk) 23:01, 18 March 2015 (UTC)
I removed the following paragraph from the History:
"While MS-DOS appeared on PC clones, true IBM computers used PC DOS, a rebranded form of MS-DOS. Ironically, the dependence on IBM-compatible hardware caused major problems for the computer industry when the original design had to be changed. For example, the original design could support no more than 640 kilobytes of memory (the 640 KB barrier), because IBM's hardware design reserved the address space above this limit for peripheral devices and ROM. Manufacturers had to develop complicated schemes (EMS and XMS, and other minor proprietary ones) to access additional memory. This limitation would not have been a problem if the original idea of interfacing with hardware through MS-DOS had endured. (However, MS-DOS was also a real mode operating system, and the Intel x86 architecture only supports up to 1 MB of memory address space in Real Mode, so for simple access to megabytes of memory, MS-DOS would have had to be rewritten to run in 80286 or 80386 Protected Mode.) Also, Microsoft originally described MS-DOS as "an operating system for Intel 8086-based microcomputers", and the 8086 CPU (and its cousin the 8088) itself has only 1 MB of total memory address space."
There is no reference in that. Since Microsoft DOS exists because IBM contracted with Microsoft for a disk operating system for their PC, it is wrong to say that PC DOS is a rebranded form of MS-DOS. Therefore it is wrong to say that IBM compatibility caused major problems. The comments about supporting no more than 640 KB of memory is irrelevant except for the fact that MS-DOS used real mode. Actually, the processor used in an IBM PC (except for the IBM PC AT and IBM PC XT/286) was incapable of using more than 1 MB of memory; "real mode" only applies to future processors. That whole paragraph was a mess. Sam Tomato (talk) 17:42, 6 May 2015 (UTC)