Talk:Windows 3.1x

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Microsoft / Windows (Rated B-class, High-importance)
WikiProject icon This article is within the scope of WikiProject Microsoft, a collaborative effort to improve the coverage of Microsoft on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.
Taskforce icon
This article is supported by WikiProject Microsoft Windows (marked as Top-importance).


Seems rather inexact ms tcp/ip was an alternative to Trumpet winsock. User:Ericd


Other than the 286-386 requirement, what were the other minimum requirements of the 3.x Windows OS? I've been trying to look into that for comparison purposes. User:SvannahLion

1MB RAM, as I recall, and about 5MB of hard drive. That was it. Ran much better with 4MB. Tannin
For Windows 3.1 my manual says: Microsoft MS-DOS version 3.1 or later. For 386 enhanced mode, a 386 processor and 640 kilobytes of conventional memory, 1 MB of extended memory and 8 megabytes of disk space.

To run Windows 3.1 standard mode: a 286 processor, 640 kilobytes of conventional memory and 256 kilobytes of extended memory and six megabytes of harddisk space.

I would suggest having at least eight MB of free harddrive space for a swapfile.

Windows for Workgroups: 386sx, MS-DOS 3.3. Three megabytes of RAM. 4 MB is recommended. With networking disabled, you need two megabyes. Ten and a half MB of harddisk space is required.

On my computer, my C:\Windows direcotry is about 13.5 MB in size. My C:\DOS directory is about 6 MB. Hope this helps.

WfW: the article is correct, the requirement for a 386 only came with version 3.11 (which only boots in extended mode); version 3.10 of WfW, apart being much less used, was able to boot in standard mode (using DOSX as extender, no VXD); of course, then, the computer cannot act as server, only netword client, since VSERVER was a VxD. So WfW 3.10 could boot on a 286 (seen on a 3MB Toshiba laptop). Also see AntoineL 18:08, 6 October 2006 (UTC)


  • Shouldn't there be something about "Microsoft Bob" in this article? DCEdwards1966 06:59, Nov 26, 2004 (UTC)
Yes, a brief mention and a link to Microsoft Bob.
-- UTSRelativity 04:09, 12 Jan 2005 (UTC)

Please Re-write[edit]

The seperate mentions of Progman and Winfile should be grouped together and put at the top part of the article. I am not sure but isn't the options bar in Windows 3.0 as well? Missing is the fact that Reversi was not included with WFW 3.11, and also not mentioned is Paintbrush. TCP/IP support was available from Microsoft for WFW 3.11 (I would not call Microsoft third-party here). I am not sure that the article should be split between Windows 3.0 and Windows 3.1x like it is, because we might as well split it up. A mention of this being the first version of Windows with the Windows Registry (though not extensively used; configuration was primarily done with INI files) is appropriate.

-- UTSRelativity 04:38, 12 Jan 2005 (UTC)

Some goofs to fix[edit]

Windows 3.1 dropped Reversi, Reversi will run in all later 3.x versions but not in 9x. It's available as a seperate download for post-3.0 Windows versions, as are some wallpaper bitmaps from 3.0 that were dropped. (Oddly, Reversi runs fine in Windows 2000 and XP, I haven't tried it with Me.)

Actually, Reversi was initially supposed to be in Windows 3.1 as well, and it still is in some beta versions of Windows 3.1, especially in build 034. Build 034 also had all the Windows 3.0 bitmaps. These things were probably only dropped from Windows 3.1 during the late beta stage, as build 061d doesn't have them anymore. - OBrasilo 10:18, 13 September 2006 (UTC)

Windows 3.11 (NOT For Workgroups) was the version to introduce 32bit Disk Access. It wasn't terribly useful since the system still had to thunk down to using the BIOS for storage operations. Other than that there wasn't much different from 3.1. This was the final version with Standard Mode for the 80286 CPU.

Actually, Windows 3.1 (NOT for Workgroups) already did have 32-bit Disk Access, in 386 Enhanced mode. NOTE: Windows 3.11 (NOT for Workgroups) only had an updated core, SETUP.EXE, SETUP.INF and some new network drivers - no other things were changed from Windows 3.1. - OBrasilo 10:18, 13 September 2006 (UTC)

Windows For Workgroups 3.11, the final 3.x version, had both 32bit Disk and File access. No special drivers were required to use them with hard drives, if the controller was directly supported by Windows. CD-ROMs and other non-hard drive storage media required 32bit compatable drivers loaded through the DOS CONFIG.SYS and/or AUTOEXEC.BAT A WFWG 3.11 system with 32megs RAM and fully 32bit compliant storage drivers was quite nice and speedy.

Sections are needed on WinG, the advanced graphics display API Microsoft was developing before dropping it and switching to DirectX, and Win32s, the 32bit API command subset that was the fruit of Microsoft's early work on designing a fully 32bit Windows system. Most programs from the early 90's that say they're compatable with Windows 3.1x and 95 are actually Win32s programs, using that add-on to run on 3.x. Microsoft incorporated all the functionality of the Win32s API into Windows 95.

Actually, some programs still came in native Windows 3.1x versions even AFTER Windows 95 was released. Good examples of that were Paint Shop Pro (Version 3.12 was the last native Windows 3.1x version, if I recall correctly) and WinZIP (Version 6.3 was the last native Windows 3.1x version). - OBrasilo 10:18, 13 September 2006 (UTC)

I don't recall ANY version of Windows 3.x (even NT 3.x) having any native ability to do anything with the right mouse button, not without 3rd party mouse drivers. (FYI, the PROGMAN and WINFILE from NT 4.5 will run in 2000 and XP, with long filename support, and they don't support any right mouse button functionality either.)

/* Paintbrush used the right mouse button. I don't know if it needed a separate driver, but i think it didn't. */

Who am I? Just someone with 22 years computer experience who has used every version of DOS from 2.1 through 6.22 and Windows from 3.0 through XP.

Don't know who left this anonymous comment but they have some of their facts wrong. 32bda was in 3.1, not 3.11; the comment about 32b compatible (note spelling) DOS drivers is vague (I know of no such limitation); there was no NT 4.5; & so on. I see no need for most of this. Liam Proven 30 June 2005 18:45 (UTC)
Yes, and I'm someone with almost 14 years computer experience, and I used every version of DOS from 1.25 through 7.10 (of the Chinese league), and Windows from 1.01 to Windows XP SP 2 (and this information is true, it's NOT sarcasm). So there, Mr. Anoynmous "22 years computer experience" Poster, it would be nice if you stopped showing off. - OBrasilo 10:18, 13 September 2006 (UTC)

Windows NT 4.5 Small Business Server and Windows NT 4.5 BackOffice Server. Search before claiming a product they made doesn't exist! To use a CD-ROM drive with 32bit disk and file access, the DOS driver had to be compatable with Windows taking control to avoid thunking to 16bit when accessing it. I ran into that many times. Fortunately the driver on the Win98SE boot disk works fine in place of most DOS CD-ROM drivers for Win 3.1x.

The product is still Windows NT 4.0. Look below for explanations on why it's so. - OBrasilo 10:19, 13 September 2006 (UTC)

Windows 3.11 needs to be mentioned in the article to differentiate it from Windows For Workgroups 3.11, especially since that version still had Standard Mode and would run on a 286.

  • Windows 3.11 (non wfwg) wasn't a product. It was Win3.1 with a patch that wasn't widely available. AFAIK, it was never slipstreamed into retail or OEM channels - msft just began shipping wfwg.
Oh really? How come I have a full disk set of Windows 3.11 (not FWG), and also in SEVERL LANGUAGES, including Thai and Turkish then? - OBrasilo 10:18, 13 September 2006 (UTC)
Same point as OBrasilo: I do have a set of 8 disks, tagged with a famous 2-letter logo, with W3.11 (French release, without the 40/128-bit mess proper of WfW3.11) inside; came with OEM licenses. Also lists it as a separate product. You are correct it was not common. AntoineL 18:16, 6 October 2006 (UTC)

Actually, 3.11 (not for workgroups) did exist, if i remember rightly it added better foreign language support but was otherwise identical to the normal 3.1, being more of a patch than a release (talk) 05:08, 21 May 2009 (UTC)

  • 4.5 on SBS and BOS are the version numbers of those products, not version numbers for Windows NT. They were NT4 with a suite of other server apps in one box.
  • You can sign your comments by typing ~~~~

SchmuckyTheCat 1 July 2005 17:32 (UTC)

Not a product?[edit]

Windows 3.11 (non wfwg) was a product, though likely a product not promoted as "new", much like Windows 95B was (to the average user) just quietly introduced with new PCs. I had an original disk set with that version number on the lables AND on the splash screen. It can also be found as a set of disk images on various (usually transitory) "warez" or "abandonware" sites. I have also run into versions with 3.1 on the splash screen but internally* identified as 3.11, so those were probably 3.1 with the 3.11 patch.

More likely it was the other way around –- see screenshot at [1]. --tyomitch 15:21, 3 September 2005 (UTC)

There was also a 3.11 Special Upgrade Edition for Windows 2.x or 3.0.

When using the right API call to get the real version instead of the 3.10 Microsoft used to avoid problems with badly written apps that were hard coded to look only for 3.10 and would see 3.11 as not "3.1 or higher".

Another clue is the different disk lables. Windows 3.11 names are this format; MSWIN3111, MSWIN3112 etc. while WFWG names are WFW311_1, WFW311_2 etc. Microsoft lists the contents of the 3.11 disks;en-us;114448#appliesto (with the disk names) but not the WFWG 3.11 disks. I dug out a set of WFWG 3.11 disks to find the naming for that version.

So there! ;) Windows (non-Workgroups) was a real, distinct (though not actively marketed/hyped) product and should be mentioned on this Wikipedia page.

It's not a 'product' but rather a 'service pack'. It just updates a few files and changes the Windows version. Mentioned now --tyomitch 15:21, 3 September 2005 (UTC)

Something else to mention in the article is the Workgroups For Windows add-on, including which versions of Windows it'd work with and the hardware requirements.

This name only has 746 Google results, and no relevant ones. If you know anything about it, please give me a link or extend the article yourself. --tyomitch 15:21, 3 September 2005 (UTC)
Even at the time there was confusion about W3.11 and WFW3.11. WFW3.11 had very good reviews, and a lot of press. And W3.11 was a tiny minor update that got no reviews and no press and inevitably got confused with WFW3.11. At the time I thought that it was a cynical exercise in re-numbering like Word 6.0, only more dishonest. (talk) 07:15, 25 January 2008 (UTC)
Cynical exercise or not, how often is Windows 3.11 described as a separate product in published materials both online and off-line? Not just reviews, especially not just reviews at the time of its release!!! I think you will find the vast majority of off-web and online published materials define Win3.11 as a separate product. Even it "technically speaking" it isn't really, that is a point for wikis other than Wikipedia. Review Wikipedia Notability Guidelines Cuvtixo (talk) 16:21, 15 August 2008 (UTC)
I agree that the vast majority of people, then and now, did not/do not realise that, unlike the WFW upgrade, the Win 3.11 upgrade was a new printer driver disk and not much else. It will be a shame if Wikipedia encourages this confusion. (talk) 05:27, 4 September 2008 (UTC)

3.11 (not for workgroups) did exist, if i remember rightly it added better foreign language support and some bugfixes etc... but was otherwise identical to the normal 3.1, being more of a patch than a release134.36.93.46 (talk) 05:09, 21 May 2009 (UTC)

WFW 3.11 Disks[edit]

You can search for WFWG 3.11 Disk Directories in Microsoft's Knowledge Base for the file listings on the WFW 3.11 disks.

IP Stack[edit]

I've seen a few attempts to change the wording to say that WfWG included a TCP/IP stack. As far as I know (and without writing my resume here, I'd know) it was required to run an add-on package. Maybe MS started shipping that add-on package in the box? Can someone point to authorative information that TCP/IP ever did ship in the retail product? SchmuckyTheCat 00:45, 11 August 2005 (UTC)

  • An anon went back to the version that said Wfwg shipped with a stack with a URL to this technet article: That's a useful article, but note the description of how to get TCP/IP onto the machine? It involves an add-on disk. It's not part of the core product. SchmuckyTheCat 02:30, 11 August 2005 (UTC)
  • The article mentions that Wolverine was released in August 1994 which provided limited TCP/IP support in Windows for Workgroups 3.11. It is correct, but I believe it can be extended a bit: there was previous TCP/IP support from Microsoft (at least as early as 1992, shipped with LanMan 2.1, to provide NetBios over IP), but the products before Wolverine were 16-bit based (so stole precious base memory), while Wolverine was VxD-based; the previous stack is said to BSD-based, and Wolverine is said to be a rewrite. I believe Wolverine should run on other 3.1x versions, but I did not check that: feel free to correct me. At the same moment there was the developement of WinSock, which emerged around that time frame and was the real thing which boosted TCP/IP over Windows. Both Wolverine and the previous 16-bit stack can drive 16-bit winsock.dll. Many of these relics (including Wolverine, as TCP32B.EXE) are still available at AntoineL 18:32, 6 October 2006 (UTC)
The only copy I ever saw was a university using the Microsoft Stack, which was a separate product. Normal people used Trumpet winsock [2]. —Preceding unsigned comment added by (talk) 06:41, 25 January 2008 (UTC)
Microsoft released a version of Internet Explorer which included their TCP/IP stack, a dialer and other software needed to connect to the internet. It worked well, if you wanted to run IE. Many ISPs licensed Trumpet Winsock and some dialer (did Trumpet include one or have it as a separate app?) in a bundle with Netscape Navigator so their customers didn't have to pay for Trumpet. AFAIK, plain vanilla Netscape would work with Mocrosoft's TCP/IP and dialer, but I recall it was a bit of a pain to change the default browser from IE if you installed Microsoft's all in one package. Bizzybody (talk) 02:32, 27 October 2011 (UTC)

MS Support[edit]

I have changed the Active State part, beacuse 1997 was the last year MS Supported Windows 3.1.

Aren't 'memory modes' in fact 'CPU modes'?[edit]

See Real mode and related articles: they all begin with ' XXX mode is an operating mode of YYY CPUs. ' --tyomitch 15:10, 3 September 2005 (UTC)

OK, I'm now editing this section. WRT this paragraph: "In either the Standard or the 386 Enhanced modes, Windows 3.1 had a functional limit of 256 MB of memory and in Windows 3.0, it is 16 MB. At the time, most 386 computers had 8 MB RAM or less, so if the memory consumed were to balloon to 256 MB, most of this would be as virtual memory on the hard disk, with massive paging slowdowns." – I don't know where it comes from. I've run WfW 3.11 with 640 MB RAM just for laugh; it takes an edit in system.ini, but runs fine otherwise. I'm removing the whole paragraph unless anyone gives me a hint confirming these limitations. --tyomitch 19:04, 9 September 2005 (UTC)

Check the About box of Program Manager. - 19:14, 30 September 2005 (UTC)
What exactly do you mean? I can't get your point. --tyomitch 09:36, 1 October 2005 (UTC)
He means that the 256 MB limitation was in Windows 3.1, and that Windows for Workgroups 3.11 didn't have it anymore. - OBrasilo 10:19, 13 September 2006 (UTC)

Operating System?[edit]

I thought the 3.x versions of Microsoft Windows were shells, not operating systems. They required MS-DOS to be able to work.

They were OSes even though they required MS-DOS. Windows provided memory management, multitasking support, and means of inter-process communication: all the tasks an OS typically implements. The underlying MS-DOS was used for just disk access and not much else. --tyomitch 09:41, 1 October 2005 (UTC)
  • They were not actually OSes, because they require an OS to run. Windows 3.x was an operating environment which runs on top of MS-DOS as a graphical user interface shell. This is why MS-DOS is required to use Windows 3.x in the first place. Also, MS-DOS actually provided the memory management, not Windows, in the HIMEM.SYS file (DOS High Memory Management Device Driver). On a final note, in the Windows template (at the bottom of this article), Windows 3.x is in the MS-DOS based section, so this also means this article should call Windows 3.x an operating environment which runs on MS-DOS instead of an operating system. — Wackymacs 20:01, 23 October 2005 (UTC)
Operating environment is defined as "GUI front-end"; Windows is far more than a GUI front-end -- see my comment above.
HIMEM.SYS just enables programs to access memory beyond the 1st MB; it's Windows to allocate it to all the running programs and to support virtual memory.
"MS-DOS based" doesn't mean that Windows isn't an OS. It only means that Windows uses portions of MS-DOS code. --tyomitch 21:45, 23 October 2005 (UTC)
Yes, but Windows 3.x are not "MS-DOS based" - they run ON MS-DOS. Windows 3.x don't use any portions of the MS-DOS code. Also, if you try to run Windows 3.x without HIMEM.SYS loaded, it will complain that HIMEM.SYS is not loaded, and return to the MS-DOS prompt. HIMEM.SYS does allocate the memory for Windows 3.x, then Windows 3.x just allocates parts of the memory, allocated to it, to other programs. And "operating environment" is NOT a GUI front-end - an operating environment is a GUI front-end + kernel, and requires a pre-existing OS to run on. Windows STARTED as Interface Manager, which was GUI front-end, but even at the 1.x stage, it became an operating environment, with programs written specifically for it. - OBrasilo 10:21, 13 September 2006 (UTC)
  • This point can be made as specious as one wants! For example, one says that Windows needs DOS (not really MS-DOS, by the way) to run. But at the same time, DOS needs BIOS to perform about anything. And the beauty of the drawing is that Windows (WfW3.11) removes about any dependancy from the BIOS! (Andrew Schulman wrote about 200 pages this about in Unauthorized Windows 95, ISBN:1-56884-169-8). Similarly, Windows 95 is a packaged product which include both MS-DOS 7.x and "Windows" 4.x, both successors of the previous versions. On the other hand, programming VxD was very specific to Windows 3.x (along with Windows/386 2.0x and Windows 9x), about nothing is related to MS-DOS in this area, even if MS-DOS is the "subject of operation". Still another point is that with Windows 3.x, in enhanced mode, you certainly can run several tasks at the same time, something which is generally difficult to achieve with DOS. And Windows 3.x runs in standard mode with the DOSX extender, which is a simple, VCPI-compatible, DPMI server, which can be replaced (althought this is not easy) by a functional similacrum, to the point to run Windows 3.x without DOS: this is exactly how Win-OS/2 worked.

I believe an operative definition is to consider an OS as a resource manager (that is how OS/360 was designed, and how I was taught OS theory); and Windows 3.x certainly manage resources (like memory, video space, even files when it comes to 32FDA in WfW3.11), and in a very different way from what is happening when using the bare MS-DOS, in real or V86 mode). However, Windows 3.x shares a large part of its programming API, which is probably the reason for the confusion. AntoineL 19:01, 6 October 2006 (UTC)

Windows is a brand name for the MS operating system. It was also a range of products over several years which started as gui available as a runtime for gui programs, competing with Gem, and slowly diverged from MSDOS, picking up it's own memory system, then it's own file system, then it's own input system one piece at time. The divergence was substantially complete with the release of Windows 98. (talk) 06:55, 25 January 2008 (UTC)
In many ways I think Windows 3.x changed the very definition of "Operating System!" There are many complicating issues-- Windows NT 3 was presented at the same time: when it was a radically different system, based on a VMS-like microkernel. It also wouldn't have been too controversial for Microsoft to call Windows 3, "the new DOS (with window manager)!" Although they didn't, presumeably for marketing reasons. Even though certain code distinguished Win from 16-bit DOS, its design was DOS-based. So it comes down to how you define an Operating System. For a similar(pointless?) controversy in another OS, review debates about whether Linux is an OS or only a kernel, and whether it should be called GNU/Linux (as GNU software tools+linux kernel) instead of just Linux. Cuvtixo (talk) 16:10, 15 August 2008 (UTC)
BTW Here's a link to article on the abovementioned Digital Research's Graphical Environment Manager-- it was not marketed as an OS, but perhaps it should have been! Cuvtixo (talk) 16:27, 15 August 2008 (UTC)
"Even though certain code distinguished Win from 16-bit DOS, its design was DOS-based." That was an intermediate stage. Windows started out as a GUI application which ran on DOS. It finished as Win98ME, a 32 bit OS which could run a 16 bit subsystem.
It doesn't just come down to how you define an operating system (or what name you call it): it also comes down to which year you are talking about. —Preceding unsigned comment added by (talk) 05:51, 4 September 2008 (UTC)
Arguments by Tyomitch and others that Windows 3.1x is an operating system are completely specious. One cannot boot Windows 3.1x or any of Windows 95's predecessors directly; MS-DOS or an equivalent operating system must be started first. Windows 3.1x is a graphical user interface or graphical shell, and nothing more, even though it includes a number of new utilities and features not provided in DOS. When Windows 3.1 is terminated, one finds oneself at -- you guessed it -- the DOS command prompt. (It's even arguable that Windows 9x releases are just graphical shells, running on top of MS-DOS 7, a 32-bit version of the earlier operating system.) —QuicksilverT @ 17:11, 6 October 2009 (UTC)
Wrong to assert that I can't boot directly to Win 3.1. I assume that when you say "one cannot boot directly", you mean that you do not know how to boot directly -- not quite the same thing. Still, sort of with you until you got to "nothing more". Unlike the understandable first error, the second error really detracts from the strength of your semantic assertions. (talk) 00:46, 1 February 2010 (UTC)
What on earth are you talking about? You CANNOT boot directly into Windows 3.1 without DOS. If your computer appears to do this, it is because someone has added WIN.COM to your AUTOEXEC.BAT file, which DOS automatically executes on start-up.SmackEater (talk) 23:39, 4 March 2011 (UTC)
You can't boot into Windows 7 without some initial help from the BIOS. Does that mean Windows 7 isn't an OS? The fact that Win 3.1 uses DOS as its loader doesn't make Win 3.1 "not an OS." Heck, even using some of the "loader's" code once it's running does not make Win 3.1 "not an OS." Jeh (talk) 02:57, 5 March 2011 (UTC)
I have the Windows 3.0 retail box right in front of me (bought back in 1992), and Microsoft never called it Operating System. Prior versions of Windows before Windows 95 were called GRAPHICAL ENVIRONMENT (you can see a small screenshot of the box here). Despite all the arguments above, it is technically wrong to call Windows 3.x an operating system. DOS was the Operating System. Sure, Windows 9x/ME were also based on DOS, but the product was marketed (emphasis on 'market', not just 'branded) as an Operating System. The install process of Windows 9x would install the underlying DOS with the Windows 9x. Windows 3.0 on the other hand was an application that could run on MS-DOS, PC-DOS, DR-DOS or even OS/2. :But I honestly don't see the need to go through all this argument. Again, simply check the retail box of Windows 3.0. Microsoft never called it an Operating System, and never marketed it as such. --Pinnecco (talk) 13:44, 15 September 2011 (UTC)
Plus if anyone has access to the computer magazines at the time of Windows 9x announcement (aka 'Chicago'), this whole issue has been covered in great detail. Moreover, Windows 3.x is to DOS what Gnome or KDE is to Linux. You might as well start saying that Gnome or KDE are operating systems? Here is a better resolution of the Windows 3.0 box. --Pinnecco (talk) 13:57, 15 September 2011 (UTC)
The Windows 3.0 article does indeed say "Operating Environment" instead of "Operating System", however this is the 3.1x article for which the box clearly includes the word Operating System on it, as well as the majority of the sources for the article. It may or may not be strictly true, but it is verifiable. --Mrmatiko (talk) 14:08, 15 September 2011 (UTC)
We're discussing Windows 3.1 here which described itself as an operating system: [3]. In any case it is easy to adopt a very simplistic position - Windows needed DOS to load, DOS is an operating system, therefore Windows can't be one. It reality it is nowhere near as clear as that. Think of it in terms of the facilities an operating system provides:
  • Memory management - done by Windows. It has often been said that eliminating the 640Kb limit is the reason Windows 3 and successors really took of when prior version were sidelined.
  • Device management. This one is a bit murky. Disk management was done by Windows if you had 32 bit file access enabled. Your sound and graphics cards were managed by Windows. It would use DOS calls to access a CD drive but they would be passed to the CD-ROM device driver rather than DOS. Networking could use either model - NetBIOS was Windows but if you were using Novell the situation was similar to the CD-ROM.
  • Process management. This one is easy, DOS had no concept of processes.
  • An API. DOS had one. Windows had its own. There is no DOS call to create a window.
The situation is not akin to Gnome or KDE (or more precisely X) to Linux. Essentially DOS is used as not much more than a bootloader. Is Linux not an operating system because it requires GRUB or similar to load it? What about Windows 95 through ME? They used DOS underneath them too. The fact that DOS was included with those versions of Windows was effectively a re-packaging job. Does something as trivial as that fundamentally alter what the system as a whole is? Crispmuncher (talk) 20:33, 16 September 2011 (UTC).
I reckon that saying "essentially DOS is used as not much more than a bootloader" is subjective when speaking about Windows 3.1x. Yes Windows 95 also came up with DOS underneath but there are significant differences when you compare it with Windows 3.1x. The most important is how both DOS and Windows 3.1 were marketed. Both were sold separately and Windows 3.1 could perfectly run under other OSes by running SETUP.EXE under PC-DOS or DR-DOS (although Microsoft tried to avoid this with some malicious code, there were no real technical limitations). Windows 3.1x works as a graphical environment as it still relies a lot on DOS for some of its I/O. Windows 95 is different because the DOS code has been modified to allow a more seamless fusion with Windows 95. Installing Windows 3.1x wouldn't modify your boot record or tamper with any OS files. -- (talk) 09:16, 20 September 2011 (UTC)

COMMAND.COM, IO.SYS and MS-DOS.SYS (or the other names of the latter two for non-MS versions) from DOS had to load first and stay in RAM for Windows 3.11 and older. AFAIK they could not be "kicked out" of RAM and leave Windows running. The only version of the 3.x series able to dispense with most BIOS calls was Windows For Workgroups 3.11 - and that only if the hardware *and* any drivers loaded through CONFIG.SYS and/or AUTOEXEC.BAT had 32 bit support for Windows to bypass BIOS calls. WFWG 3.11 had two separate checkboxes, one for 32 bit disk access and one for 32 bit file access. Windows would detect if either or both or neither was supported and would grey out one or both options if they were not. It was pointless to only turn on one because of speed hits due to having to thunk between 32 bit mode and 16 bit BIOS access. But anyway, since all the core parts of DOS had to load first and stay in RAM, Windows 3.x is a graphical operating environment running on top of DOS. What could be changed is the Shell program. The default was PROGMAN.EXE but there were several 3rd party shells, some a single app and some meant to completely replace PROGMAN.EXE as the environment. It was also possible to replace the shell with a single task program like a web browser. Doing that locked out access to the rest of the file system as long as that app had no file system access. I recall there's a way to setup the last 16 bit version of Internet Explorer in a kiosk mode as the shell program. The reason for doing this was when the shell was quit, Windows itself would quit. Another part of a kiosk setup with Win 3.1x was to have WIN.COM automatically relaunch if it quit. (Shell replacements are also possible with Windows 95 and 98. I did that once with a 68K Mac emulator on 95b on a 75Mhz Pentium Dell to make it appear to be booting Mac OS 8.1.) Bizzybody (talk) 02:25, 27 October 2011 (UTC)

COMMAND.COM could be unloaded even under DOS applications if need be. If you ever used a floppy installtion you'd occasionally be asked to re-insert the boot disk after exiting an application so that it could be reinvoked. Crispmuncher (talk) 15:43, 27 October 2011 (UTC).

I'm sorry for rocking the boat a little, guys. I came along to clarify the intro a bit. But unfortunately, its most wonderfully accurate prose was stricken down by a later editor, in the interest of explicitly precise citations. Maybe someone can find the citation to concretely justify restoring it to its former correctness. It had said this: Windows 3.1x is a series of 16-bit operating system shells, marketed as operating systems Smuckola (talk) 10:36, 20 March 2013 (UTC)

Indeed, can we stop the "this is an operating system" silliness? I just effortlessly installed and started Windows 3.1 in DosBox, of all places (running in full "enhanced" mode too as others had mentioned it could), and took another look at the original README.WRI. Section 1.0 of the same reads as follows (all emphasis mine):
1.0 Running Windows with an Operating System Other Than MS-DOS
Microsoft Windows and MS-DOS work together as an integrated system. They were designed together [...] . 
Running Windows version 3.1 on an operating system other than MS-DOS could cause unexpected results or
poor performance.
Not only can this "operating system" trivially be a layer above several completely independent programs that are perfectly usable systems in their own right, but even the official documentation Windows 3.1 ships with does not once explicitly refer to it as an operating system (no, not in any of the other chapters either). I hope that's enough. No qwach macken (talk) 21:04, 25 March 2013 (UTC) A picture for the nonbelievers among you. Also note that much of the discussion up to now consists of random opinionating. No qwach macken (talk) 21:20, 25 March 2013 (UTC)
Hello guys. Faith, truth and analysis are valuable but the Wikipedia policy is Wikipedia:Verifiability. That means one should cite a reliable source like this or this that says Windows 3.1 is an operating system. (Or maybe a reliable source that says something else.) Synthesis of published material that advances a position, like what User:No qwach macken did above, is not allowed. Best regards, Codename Lisa (talk) 22:09, 25 March 2013 (UTC)

There are a couple of sources saying its a graphical environment rather than an operating system atleast we should mention it in the article 1 2 3 4(hints so) 5 6 7 8,9 10 An introduction to Microsoft Windows 3.1 June Jamrich Parsons

11 Windows 3.1 Fundamentals Carolyn Z. Gillay

13 New perspectives on Microsoft Windows 2000 MS-DOS command line: comprehensive Harry L. Phillips, Eric Skagerberg


15 sort of

Catverine (talk) —Preceding undated comment added 21:42, 20 November 2013 (UTC)

Full OS or MS-DOS shell?[edit]

(See additional discussion at Talk:Windows_3.1x#Operating_System.3F, above.)

That headline goes unanswered I suppose... —The preceding unsigned comment was added by (talk) 13:00, 31 March 2007 (UTC).

This section applies to several versions of Windows, so shouldn't it be in a more version-neutral article than Windows 3.1x? Josh the Nerd 00:31, 7 July 2007 (UTC)

All versions of Windows are "something that runs on something", in the sense that there is no unique Win32 or Win16 boot loader. None the same, once it is booted from DOS, it takes over the computer, and DOS becomes a real-mode driver that happens to get fed whatever Windows needs to feed it. Still, Netware and Deskview do the same thing: Netware boots from DR-DOS, and then proceeds to loads drivers etc that DOS can not run.
Unlike OS/2, and i presume, Linux, the text-mode setup of Windows NT is a different beast to the command prompt. Very few applications will run here (except for some specially written apps like O&O Bluecon, CmdCons).
--Wendy.krieger (talk) 14:18, 10 January 2009 (UTC)
The discussion up there mentioned that Windows 3.0 is an "operating environment" and was marketed as such, while Windows 3.1 is an OS. But the very name "Windows 3.1" implies a mere incremental improvement over the previous version. I was under the impression that the whole reason Microsoft changed the format of the name with Windows 95 was that they were making the transition to a true Windows OS, and thus wanted to make sure it wouldn't be confused for merely an enhanced version of Windows 3.1. (talk) 10:19, 14 January 2012 (UTC)

Operating system or GUI?[edit]

Is it proper to call Windows 3.1 an operating system? The Windows 3.0 calls it a Graphical user interface. Windows through 3.1 and 95, 98, and Me used DOS as the o.s. Windows 2.0 calls it an Operating environment. Bubba73 You talkin' to me? 20:18, 4 November 2012 (UTC)

Yes, I think it is proper to call it an OS. While it did depend on a DOS implementation to start, and allowed you to stop and drop back into that same DOS, it also added e.g. (a limited form of) memory protection, shared binaries (DLL's) that could be used by multiple processes (or tasks, as they were referred to back then IIRC). I'm not 100% sure about 3.10, but I do know that 3.11 got both native disk cache and even a TCP-IP stack (both 32-bit even, if memory serves), and a lot more than any DOS by itself could do. While it did have fallback paths like mscdex.exe (?) driving CD-ROM drives at the time, most of what it did was self-contained, and the parts that weren't were mostly (only?) to provide compatibility with DOS and the large amount of DOS programs able to run in the "DOS-Box" - and of course the ability to terminate itself and return to that DOS mode. But as DOS was mainly for bootstrapping, much like e.g. Linux, NT, or just about any system needs a boot-loader too, I'm very much inclined calling Windows 3.1x an Operating System. 9x discussions do not apply (except to say they definitely didn't use DOS as their OS). (talk) 15:48, 19 November 2012 (UTC)

The Windows 95 article says "Windows 95 integrated Microsoft's formerly separate MS-DOS and Windows products." You couldn't run Windows 3.1 without DOS, so it seems to me that DOS was the OS and Windows was the interface. Bubba73 You talkin' to me? 16:15, 19 November 2012 (UTC)
Windows 3.1 replaced the better part of DOS in memory and took over most device handling. For all intents and purposes, it is an operating system. Edokter (talk) — 23:14, 23 November 2012 (UTC)
The article says "...which typically came with DOS 6.22 on one CD". Why did it come with DOS if it didn't need DOS? Bubba73 You talkin' to me? 23:38, 23 November 2012 (UTC)
I don't know if this is a good reference, but this says:

Windows 3.1 is not entirely a new operating system, it is a graphical interface that is "built over" MS-DOS. This type of operating system is referred to as a DOS "Shell" utility. Windows 3.1 translates a user's point-and-click instructions into DOS commands for DOS to execute. Any request that is entered through the Windows 3.1 graphical user interface shell will inevitably be performed by the MS-DOS operating system.

Bubba73 You talkin' to me? 23:45, 23 November 2012 (UTC)
And see Operating system shell. Bubba73 You talkin' to me? 00:13, 24 November 2012 (UTC)
Windows started out as an operating environment rather then a mere shell. By the time WfW 3.11 came out, all DOS/hardware function were virtualized, and DOS itself was only used as a bootstrap, much like Windows 95; it did not run "on top" of DOS, it completely replaced it in memory. The fact that DOS was needed to load Windows is not the main factor here. Edokter (talk) — 10:38, 24 November 2012 (UTC)
And yet, Windows 3.1 can run entirely inside a DOSBOX window on a 32 bit OS (indeed, that's the only way to get it to work beyond VGA mode with today's graphics adaptors), and when you look at it that way, it certainly seems like nothing more than a fancy graphical shell for MS-DOS Linuxgal (talk) 21:41, 30 December 2012 (UTC)
It could do so only in 16-bit 'Standards mode' (which even WfW 3.11 no longer had), definitely not in 32-bit '386-enhanced mode', as the host OS already has taken control of the CPU's virtual addressing capabilities. You really need to know the inner workings of Windows a bit more before stating such poor arguments. Edokter (talk) — 22:53, 30 December 2012 (UTC)
On the contrary, it not only runs 386-Enhanced under DosBox, it can run in multiple instances of DosBox. I'm not making an idle boast, or repeating something I think I remember from long ago, here's a screenshot from last October. Linuxgal (talk) 00:40, 31 December 2012 (UTC)

Split proposal[edit]

I think that the article should be split into Windows 3.0 and Windows 3.1 (or Windows 3.1x again, whence it was moved on January 11, 2003), because most links to here mean either of them two, and not both. Also there's comparatively little content now that applies to both of them. Cf. older comments:

This article seems to be more about Windows 3.x than 3.1x specifically. Should it be renamed? --Brion 21:57 Jan 10, 2003 (UTC)
I've usually seen 3.x, so that's probably a better name for it. (There was a Windows 3.0, also) -- Wapcaplet

--tyomitch 13:12, 13 November 2005 (UTC)

History from MS, 1.01 through 3.11[edit];EN-GB;q32905

Details on 1.01 through 3.11 Standard Edition. 3.11 (NOT For Workgroups) was definately a seperate, complete product. I encountered it on many OEM PCs made between 1993 and 1995.

I'm still hunting for info specific to the Workgroups for Windows add-on package. I saw it in many computer catalogs in the 93 to 95 years, right next to the listings for WFWG 3.11. Anyone have an old Computer Shopper from back then? :)

3.11 12/31/93 Requirements

                    - Same as version 3.1
                    - Certificate of Authenticity
                    - More sophisticated hologram and an MS (3M) sticker
                      on box
                    - An 800 number to call (in the United States &
                      Canada) and check for product legitimacy
                    - Updated drivers
                    - Five updated core files
                    - NetWare support files (from Novell)

Win310 was the first package, both OEM and upgrade.
Wfw310 followed. It was very slow, poorly marketed, and was dubbed "Windows for Warehouses" (part of the Windows Everywhere schema). It's kind of like Win310 with Networking tools.
Wfw311 appeared. Many of the problems that plagued Wfw310 had beem fixed, and it ran faster than Win310 on similar hardware. Microsoft pushed it hard as replacement for both OEM and upgrade. It should be noted here, as with DOS, that the intent was to distance Windows from OS/2. DOS 6.x included the OS2.TXT file (how to delete OS/2 if found), and Windows 3.1x included some new features that were incompatable with the then current OS/2 2.1x for Windows. IBM worked around these!
Win311 appeared as an OEM version. It's pretty much Win310, with drivers updated, and a new setup to handle the better compression given by compress.exe v1. (KWAJ). The version has something to do with Microsoft trying to comandeer the then-active market for peer-to-peer networking (eg netware lite, baynan vines, etc).
Wfw311ao appeared. It is, like the DOS Step-Up disks meant to replace an earlier version of W31x. It's a complete version of Wfw311, with a hacked setup. Disks 2-8 are indeed identical.
The Windows 3,51 server disk includes Wfw3.11 setup and patches (eg tcp/ip, y2k). However, none of these patches are slipstreamed, it's exactly as if you install Wfw 3.11 from the original diskettes, and then install the various patches.

--Wendy.krieger (talk) 14:05, 10 January 2009 (UTC)

But will Wfw311ao (Workgroups for Windows 3.11) install on a 286 running Windows 3.1 or 3.11 in Standard Mode? Or does it upgrade to Windows For Workgroups 3.11 which only runs in 386 Enhanced Mode? If it does that, does it also add 32 bit disk access and 32 bit file access? Bizzybody (talk) 02:05, 27 October 2011 (UTC)

Windows 3.2[edit]

I corrected the short paragraph on Windows 3.2 and added a screenshot. It's based on wfwg, it includes all the stuff you expect from wfwg like the ms-mail client and networking add-ons (some of that isn't localized either). I didn't use these two images, but if anyone likes them better than the splash screen they could be swapped out. SchmuckyTheCat 15:07, 11 April 2006 (UTC)

Excuse me? I have Windows 3.2 on my hard disk and it doesn't have any WFW features. Actually, there has been Simplified Chinese Windows 3.1 released in the 1993, and Windows 3.2 is simply an updated version of that, with some different icons and new and improved IME's (Input Method Editors). Source for this information: my testing of both version, and this article:;en-us;129451 , which as you can see, comes from Microsoft themselves.

I think that the reason why you thought it was based on Windows FWG 3.1 or 3.11 is that you have probably installed Windows 3.2 over Windows FWG 3.1 or 3.11, so of course, the Windows FWG 3.1 or 3.11 additional programs were preserved. I think it's kinda obviuos, since you yourself clearly said that some of the stuff you have in Windows 3.2 is not localized - if it was actually part of Windows 3.2, it would have been localized. - OBrasilo 17:45, 11 September 2006 (UTC)

My install was not an update. I installed it clean simply to take those screenshots. SchmuckyTheCat 18:39, 11 September 2006 (UTC)
Then please make some screenshots of MS Mail and the Network Add-Ons that are supposed be in Windows 3.2. If you manage to do that, and those features really are there, then maybe what you have is a rare Windows for Workgroups 3.2 edition. - OBrasilo 17:02, 12 September 2006 (UTC)
If the images are Fair use we really can't use them on the talk page, linking to them works, so I changed them to a link. PPGMD 02:46, 9 October 2006 (UTC)
All screenshots of MS software are used with permission, not fair use. There is no prescription against using them on a talk page. SchmuckyTheCat 05:34, 9 October 2006 (UTC)
That cannot be true. As a big DOS/Win3.x fan from China DOS Union, I frequently play around with DOS and Windows 3.2 (Chinese version of course). The WFW features are NOT included in Windows 3.2. There exists a seperate Chinese version of Windows for Workgroups 3.11 which I have also installed, which is distinct from Windows 3.2.--Wengier 02:43, 14 April 2007 (UTC)

Three one one[edit]

I've read through the above discussion, and I side with SchmuckyTheCat against the man who didn't sign his comments. The bit where he says "Who am I? Just someone with 22 years of computer experience etc" angered me with its naked display of self-righteous narcissism. However, I'm confused by the following text in the article:

Windows 3.11 was not a standalone version of Windows, but rather an update from Windows 3.1, much like modern Windows service packs. For those who did not own Windows 3.1, full disk sets of Windows 3.11 were sold.

It seems to contradict itself. Does it mean to say that "for those who did own Windows 3.1, full disk sets of Windows 3.11 were sold", or does it mean to say that "later editions of Windows 3.1 were bundled with the 3.11 upgrade as an integral component", or something else? I would like to know this so that when I rewrite that paragraph it is not misleading. Lupine Proletariat 10:56, 18 May 2006 (UTC)

Actually, Windows 3.11 (non FWG) was a separate product. Of course, there was the update patch, but it didn't update everything. For instance, the full product also had an updated SETUP.EXE file (and until Windows FWG 3.11, SETUP.EXE had both the DOS and the Windows parts of Setup, so it was indeed installed into the Windows directory as well), and some other files as well, including SETUP.INF. If I'm not wrong, Windows 3.11 (non FWG) even had some updated network drivers or something, but I'm not sure. I'll check that when I come back home (should be on Friday). Windows 3.11 (non FWG) was also released in several languages, including Thai and Turkish. - OBrasilo 17:58, 11 September 2006 (UTC)

3.11 (non FWG) in the full package included either 32 bit disk access *or* 32 bit file access (been a long time since I've used it!) if run on a 80386 or higher CPU. But since it only had one of those, enabling it was pointless because it still had to thunk between 16 and 32 bit during file operations. The only benefit came with WFWG 3.11 that had both 32 bit disk and file access - but the hardware and all DOS drivers for storage loaded before Windows had to support both. With both enabled it made things faster because Windows could bypass the BIOS for file operations. If your PC couldn't support both, one or both of the checkboxes to enable them in windows was greyed out. I ran 3.11 on a 286 with 12 megs RAM (two four meg Micron and one Everex three meg ISA boards full of DIP chips plus one meg on motherboard). I upgraded to WFWG 3.11 with my first 386 but I do remember wishing I had a 386 with 3.11 so I could enable that lone 32 bit checkbox to see if it would go faster. What I do wonder is if the 3.1 to 3.11 update did *not* include that small bit of early 32 bitness? Bizzybody (talk) 01:34, 27 October 2011 (UTC)
Hazy memories here too, but that does not ring true. From memory Windows 3.0 introduced 32-bit disk access (swapfile only) which was supplemented with file access (everything else) in 3.1, and yes, both at the same time. I do have a Windows 3.1 VM here that I may fire up and check. As a side note, something isn't quite right about that memory either: memory on ISA boards would be completely useless to Windows which has always used extended and not expanded memory. Crispmuncher (talk) 01:48, 27 October 2011 (UTC).

Janus merge[edit]

I support the proposed merger. There does not seem to be enough information (yet) to justify a separate article. -- Seitz 04:20, 8 July 2006 (UTC)

Errr... there is actually a lot of information and even screenshots about Janus. I can provide screenshots of builds 034 and 061d (Final Beta Release), as well as of the DOS portion of the Setup of build 068 (Pre-Release Build), and I can also provide a detailed description of the so-called AARD Detection Code, which has been incorporated into some files in build 061d, and remained until Windows FWG 3.11, when it was finally removed (although it was disabled in the final version of Windows 3.1 and all subsequent Windows 3.1x versions, up to Windows FWG 3.11, but including Windows 3.2).

I can also provide a detailed description of those 3 builds, as well as how they differ among each other, from Windows 3.0, and from the final version of Windows 3.1.

So, summarized, I think that a separate article for Janus is more than justified. - OBrasilo 18:04, 11 September 2006 (UTC)

Actually, I was wrong back then. Janus was not the codename of Windows 3.1. — Preceding unsigned comment added by OBrasilo (talkcontribs) 18:04, 19 March 2014 (UTC)


This sentence is extremely unclear: "Pre-NT Windows systems, not only 3.x and earlier but also 95, 98 and ME, have a complex, original, hybrid and not fully documented internal structure that differ across modes, which was, in order of more complexity and more like operating systems and less DOS use, real mode, standard mode, and 386 enhanced mode." Normally, I would probably just re-write the sentence myself, but in this case, I cannot figure out what it is even trying to say. --Miken2005 19:48, 26 July 2006 (UTC)

It seems the sentence I'm complaining about was edited by on 11 July 2006. Since the version prior to then is clearer, and no other edits have been made since then, I've reverted the article to its previous version. If the bit added by is significant, please re-write it in a clear fashion. --Miken2005 20:02, 26 July 2006 (UTC)

How about "The Windows that descended from 1,0 (including 95,9 98, ME), booted from DOS, and used DOS as a device driver. The DOS/Windows interface was not documented, and various transgressions to published interfaces were made in the line of efficency. This served also to make it harder for other OS manufacturers, like DRI and IBM, to adjust their operating systems to run Windows. The increasing amounts of memory lead to a transition from real-mode operations (Windows 1.0 is a shell, Windows 2.x and later are operating systems), to protected mode (in the form of standard and enhanced). Since the more advanced memory models require more system memory to run, one would tend to boot windows into a mode that was more compatable with installed memory and the application at hand. Real mode was last used in Windows 3.0, Standard mode was removed in Windows for Workgroups 3.11 (Windows 3.11 had it).
A new kernel was developed in the form of NT, designed to replace DOS and its limitations. This appeared in Windows NT 3.10, however, it was not until 2002 that hardware became common enough for Microsoft to switch users from DOS-based Windows to NT-based windows (in the shape of Windows XP). --Wendy.krieger (talk) 10:38, 11 January 2009 (UTC)

Delay to release of Windows 3.1 due to coders' obsession with Lemmings for the Acorn Archimedes[edit]

Can anyone (or does anyone know of anyone) who can personally corroborate this?: --Sctb 20:17, 02 August 2006 (UTC)

To the 'Citation needed' comment for the release date[edit]

The release date of the current version (Win 3.11 for Workgroups) is 11/1/1993 as per Someone please update it and give the citation. Shijaz 05:12, 16 July 2007 (UTC)

Fair use rationale for Image:Windows 3.1.png[edit]

Nuvola apps important.svg

Image:Windows 3.1.png is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 21:13, 24 October 2007 (UTC)

Fair use rationale for Image:Windows Logo (horz).svg[edit]

Nuvola apps important.svg

Image:Windows Logo (horz).svg is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 21:34, 24 October 2007 (UTC)

Fair use rationale for Image:Windows 3.1.png[edit]

Nuvola apps important.svg

Image:Windows 3.1.png is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot (talk) 20:04, 26 November 2007 (UTC)

Windows 3.11 still supported[edit]

Windows 3.11 embedded is being withdrawn from the market November, 2008. This would indicate that Windows 3.11 is, technically, still supported. —Preceding unsigned comment added by (talk) 18:35, 11 July 2008 (UTC)

Not necessarily supported, just available. - Josh (talk | contribs) 18:44, 11 July 2008 (UTC)

Forth version of Windows?[edit]

The very first sentence of this article states this to be the forth major release of Windows, but is that really true? I see a number of problems with the statement. For a start the page groups two distinct versions of Windows together: how can both be the forth release? Next does the jump 3.0->3.1 really qualify as a major release? It is only a minor bump to the version number and less significant than 95->98, for instance. It is probably more akin to 95 (original release) to 95 OSR2. Finally what about Windows/386? I would argue that that was a major Windows release even if it didn't have a version number associated with it? If no one has any comments on this I am minded to rework the opening of this article to remove the issue. CrispMuncher (talk) 19:15, 11 July 2008 (UTC)

Go ahead and re-work it to remove the issue, but 3.0 to 3.1 was a major upgrade. More than 95->98. Big kernel changes are a big deal, even when users don't see it. SchmuckyTheCat (talk)
It's hard to imagine it as the 'fourth version', since there is 1.00 to 1.04, 2.0 to 2.03, 2.10, 2.11, (2.1x in 286 and 386 versions), 3.0, 3.0a, 3.0 MME, 3.10, 3.10 Wfw, 3.11, 3.11 Wfw, 95, OSR 2.0, OSR 2.1, OSR 2.5, 98, 98SE, ME. Excluding the various bug fixes, one can say, eg 1.x, 2.0x, 2.1x, 3.0x, 3.10, 3.11, 95, OSR/2, 98, 98SE, and ME are the major versions. Different things appeared at different places: for example, DOS in 95 - ME is essentially identical, except for the version number, and that the kernel is upgraded to support fat32 (7.1), and then as a Windows loader (8.0). Windows 3.x do strange things to DOS, (even 3.0, which is why winoa20.386 exists). MME extentions appeared in versions of 3.0, standard in 3.1; Networking in 3.10, standard in 95; msdos.exe as shell to 2.1x, supplied with 3.0, etc. --Wendy.krieger (talk) 11:16, 11 January 2009 (UTC)

AARD code[edit]

I have just been reviewing some edits made by user Cuvtixo and I have some doubts about them, but on the other hand I am reluctant to simply revert what are obviously good faith edits without consultation.

Firstly the new title is itself misleading: there was nothing misleading about the error - it was simply vague, but even then not unusally so by the standards of the time. This is potentially POV since it casts MS in an unduly negative light (the whole AARD thing does anyway, but we still need to keep a sense of perspective here).

Secondly the ordering of this revised section leaves a lot to be desired. It says that the tests were disabled in one breath and then continues to the problems it causes. This means the article does not flow very well - do the problems desribed manifest themselves after the code had been disabled? The article isn't clear there at all.

Finally, we already have a dedicated page on this subject, so do we really need to go into so much detail here? The AARD code is of comparatively niche interest since it only affected a small proportion of users and even then was easily circumvented. Having too much coverage here distract from the wider article as a whole. CrispMuncher (talk) 14:24, 15 August 2008 (UTC)

Thanks for posting before reverting!!! I understand you're point of view about the word "misleading" which is why I included a question mark at the end. On the other hand, the previous title "AARD code" is too vague. One might ironically say the section was "obfuscated!!!" haha. I do need to verify the Brad Silverberg quote: "later sent another memo, stating that "What the [user] is supposed to do is feel uncomfortable, and when he has bugs, suspect that the problem is DR-DOS and then go out to buy MS-DOS" The original source is invalid. The warning message can apparently manifest, but it needs to be triggered by device drivers. I left this out because, as you say, there is a dedicated article (which I consulted as I edited). The warning message has NO EFFECT WHATSOEVER-- This is not, and was never a user issue!!-- But it was significant enough for Caldera to bring a lawsuit carried over from Novell, and I think of MORE significance because it was not a technical issue. I was considering adding a section exploring the fact that the code was obfuscated-- It certainly leads many to think it was done as part of a deliberate attempt to undermine competitors, and "sets the stage" for accusations of unfair anti-competitive practices in the anti-trust convictions Microsoft would eventually get.
I think the AARD section is important because it addresses 1.history of this software's development 2. Microsoft's business history 3.The possibility of running (strictly speaking, booting) Windows 3.x on DOS other than PCDOS or MSDOS, not otherwise adressed, etc, etc. Cuvtixo (talk) 15:17, 15 August 2008 (UTC)
There are many citations given, which is evidence of the subjects' significance. I would think time spent on finding citations for the previous section should be more of a priority than making this section "flow"!!!!! Cuvtixo (talk) 15:25, 15 August 2008 (UTC)
I am certainly no Microsoft apologist - I steer clear of Windows and MS where ever possible and have been using Unix systems since before Linux was even conceived. Yes, the AARD code does show MS in a negative light and it is right that it do so. However, this has to be kept in proportion - we can't simply say "since this makes MS look bad any other criticisms can be thrown in no matter how ill-founded" - that would be just as POV as attempting to gloss over the issue. Describing the AARD error as misleading falls into this category - it was not telling the user any falsehood, it simply warned that it was not running atop of MSDOS. You can argue the differences between MSDOS and DRDOS were sufficently minor that this was an irrelevance, and the documents released as part of the antitrust case show that the motivation for the warning was not technical, but this does not in and of itself make the error misleading. Personally I see nothing wrong with "AARD controvesy" as the subheading but I can see your point and I'm quite willing to concede it, provided it is changed to something that is both neutral and accurate - "Compatibility with DR-DOS", perhaps?
I have no concerns about the notability of the AARD code - it is certainly something that we should be covering. However, we do have a dedicated page on it so we don't need to cover it in tremendous depth here. The previous version gave a brief outline of the issue and there was a link to the relevant page for more details if desired. There seems to be a general tendency on Wikipedia to define the scope of the article too broadly leading to mass redundancy between articles. If we start describing the AARD controvesy in detail here then we risk losing the main focus of the article - Windows 3.1x, as we describe MS-DOS, DR-DOS and successors (and its various owners), the antitrust settlement and the Caldera legal action, none of which has anything other than the most tenuous relationship with the subject in hand. Reference AARD in summary here - further details are on the AARD page if desired.
I disagree with your views about article flow though. It doesn't matter how well sourced each statement is, if each statement is freestanding and they are inserted into a document at random you end up with a jumbled mess of statements. A little editorialising is needed to assemble those statements into a coherent article that flows from the background information to the consequences and ultimately on to any conclusions. I must declare a interest here since I did a certain amount of refactoring of the AARD page earlier this month. Compare that article before and after. JeremyMcCracken did sterling work expanding the article, finding references and generally adding content. However, essential details were introduced at the last minute and as such the details of the lawsuit were lost amid details that should have been introduced at a more logical point earlier in the article. That article was restructured so the background was presented at the appropriate time and the lawsuit could be explained without excess chaff distracting from the issue. Within the scope of a separate article, we have the space to present details in a logical manner like this - attempting to shoehorn everything into a few sentences as one section of a larger article makes things much more diffcult.
Finally, consider this from a user perspective. "Windows 3.1x" is a fairly vague topic to search for and it is likely that a user looking at this page is basically after an executive summary of these particular Windows versions rather than an exhaustive account of each and every facet. If memory serves it was these versions of Windows that introduced OLE 2.0. This page doesn't even mention OLE, which it probably should, but do we need to describe it right down to the API level here, or can we simply say that this is a specialist area better covered elsewhere? If we covered it and all the other similar areas comprehensively the page would run to thousands of pages of A4, and be little use as a reference as the user would be unable to sort the wood from the trees. Hyperlinking between articles is the biggest single benefit of an electronic encyclopedia and whilst each article should be self contained, it isn't necessary for each page to effectively devolve into "The sum total of human knowledge". CrispMuncher (talk) 20:53, 16 August 2008 (UTC)
You've given a lot of thought to this! I'll try to address your points in reverse. I disagree with your version of the "user perspective." I happen to know a bit about the story of Symbolics Lisp Machines and I might have assumed Wikipeida "users" would generally be looking at its aspects as a highly advanced system from the LISP AI heyday. Actually, many "users" look at Symbolics as a business case study, where MIT engineers with much hailed and revered products and great private and gov't backing still failed in the marketplace. It is also the company that inspired Richard Stallman to reject proprietary software, as he worked to match Symbolic's features on Lisp Machine, Inc.'s (a legendary feat in hacker circles). I think this article already contains much about "technical" aspects of Win, and not nearly a balance of the social impacts of the product. Windows 3.1x defined "Operating System" for the general population, despite being a strange kluge of DOS and NT tech, and despite little consumer exposure to Windows 1.x-2.x It was THE best selling software ever, and really wrested control of the PC marketplace away from hardware companies (especially IBM's PC division) to a software company. From an Economics view, nothing like it has happened before or since. IBM, Sun, HP, Digital and even Apple Inc, only offered computer+OS combinations which could never be as profitable. Thinking ten, twenty, thirty years from now, these sorts of details will be much more important to the Wikipedia user, and I'm confident that Wikipedia will still be read then, at least in some form. The AARD code is a small but I think important note of Window's social and economic importance, putting the Win 3 product in historical/sociological context. Perhaps there are better ways of doing this, but I would compare/contrast this article to the Mac OS article, which refers to lots of non-specific-Mac-OS info about AU/X, the Star Trek project, Mac OSX and various emulators, without running into thousands of pages (Thousands of A4s? Why not a comparison of American legal length pages?!) Cuvtixo (talk) 20:04, 30 August 2008 (UTC)
Sorry, meant to get back to this somewhat more promptly than I have. This conversation has got fairly abstract and if we are to get anywhere we need to return to some concrete proposals. I've made some alterations now and I'll briefly summarise them here:
  • I've changed the section heading to "DR-DOS compatibility" so that is is more neutral. It is inappropriate to describe a numeric error message as misleading as I have already discussed.
  • I've added a Main article link at the top of the section - we only need a summary here.
  • The actual section has been trimmed down so that it does not reference too much information that is extraneous to Windows 3.1x.
  • The previous version was a mess in respect of ordering - it did not follow any logical sequence, either chronological or by argument. It is now clearer which statement follows on from which - the previous version created the impression that the installer failed after the test had been disabled.
  • I've made the hyphenation of MS-DOS and DR-DOS consistent throughout the section.
  • Name of the lawsuit trimmed. Does "Caldera vs Microsoft" really add anything when we have already named the two parties involved?
CrispMuncher (talk) 21:41, 6 October 2008 (UTC)

Windows 3.x retired[edit]

So according to it was on 1 November 2008 that Windows 3.x was retired. Until then Microsoft issued licences for the program. This could go into the article, maybe? --Joshua Issac (talk) 01:17, 5 November 2008 (UTC)


I tried different variations on the broken infobox which I found at the top but can't make it work. Maybe the OS version infobox doesn't exist? James Skinsale (talk) 07:07, 4 June 2009 (UTC)

Addons: WinG[edit]


I remember there was another important addon library for 3.11. I think it was called WinG and can be considered the precedessor of DirectX. It needed the Win32 extension to run and I think only very few games used it back then.

See here:

The link to the Microsoft FTP still works :)

 —Preceding unsigned comment added by (talk) 16:06, 26 June 2010 (UTC) 

What about the registry?[edit]

The Windows Registry, the controversial configuration database either lauded or accursed by developers and users alike, began with Windows 3.1x. There's no mention of it here, even though there's a rather good article elsewhere on Wikipedia about it. Could someone fix this? (talk) 23:44, 16 December 2012 (UTC)

I added a sentence to the lead section, because I didn't see a better place to put it. Bubba73 You talkin' to me? 01:24, 17 December 2012 (UTC)
Windows Registry actually started with MS-Office, and migrated into Windows. If you get a hold of the original first-day release Windows 3.10 diskettes, from 1992, they already have office registry keys in Windows 3,10, and an office install under 3.00 will create a registry. Also regedit is one of the proggies that runs under 3,00 for that reason. Wendy.krieger (talk) 13:33, 10 June 2013 (UTC)