Talk:Timeline of DOS operating systems

Windows 98
no entry about win98?188.222.7.234 (talk) 09:46, 15 July 2011 (UTC)
 * As far as I know, Windows 98 did not introduce a new version of MS-DOS. I believe it came with the same version of DOS as Windows 95 OSR 2.  So, from the standpoint of MS-DOS, Windows 98 may not be particularly meaningful.  Though I don't think there would be any harm in including Windows 98 in the timeline, users such as Codegen86 (see below) may disagree.  I am curious about how these Windows versions of DOS "identify" themselves as "MS-DOS version 7.x".  Whenever I do a VERSION command, I just get "Windows 95" or "Windows 98" etc., not MS-DOS.  Can anyone tell me how to find the MS-DOS version numbers which are embedded in Windows 9x?  Thanks  Wbm1058 (talk) 13:52, 29 November 2011 (UTC)
 * I lifted the following info from some old notes of mine examining version APIs and internal file versions:
 * - Windows 95: MS-DOS 7.00 (4.00.0950) + Windows 4.00 (4.00.0950)
 * - Windows 95 A OSR 1: MS-DOS 7.00 (4.00.0950) + Windows 4.00 (4.00.0951)
 * - Windows 95 B OSR 2: MS-DOS 7.10 (4.00.1111) + Windows 4.00 (4.00.1111)
 * - Windows 95 B OSR 2.1: MS-DOS 7.10 (4.00.1111) + Windows 4.?? (4.03.1212)
 * - Windows 95 C OSR 2.5: MS-DOS 7.10 (4.00.1111) + Windows 4.?? (4.03.1214)
 * - Windows 95 OPK 3: MS-DOS 7.10 + Windows 4.10
 * - Windows 98: MS-DOS 7.10 (4.10.1998) + Windows 4.10 (4.10.1998)
 * - Windows 98 SP 1: MS-DOS 7.10 + Windows 4.10 (4.10.2000)
 * - Windows 98 SE: MS-DOS 7.10 (4.10.2222) + Windows 4.10 (4.10.2222)
 * - Windows ME: MS-DOS 8.00 (4.90.3000) + Windows 4.90 (4.90.3000)
 * The list is known to be incomplete (f.e., I did record only the DOS versions, not the DOS revisions). --Matthiaspaul (talk) 15:22, 4 July 2012 (UTC)

Title
The title is completely misleading. The timeline is not restricted to DOS, and it starts way before any x86 systems appeared. Codegen86 (talk) 15:19, 19 November 2011 (UTC)
 * I disagree that the title is "completely misleading", though I am largely responsible for expanding it from its original focus. The portion in the 1970s before the release of version 1.0 lays out both the hardware and software foundations on which DOS was built.  Without these prior developments, DOS would never have happened.  Likewise, hardware and related software developments in the 1980s and early 1990s relate closely to and often control the developments in DOS itself.  Feel free to suggest a better title.  Would "Timeline of the x86 DOS era" work better as a title?  Wbm1058 (talk) 13:52, 29 November 2011 (UTC)
 * How about simply "Timeline of the DOS era"? That implies x86 while not excluding mentions of non-x86 (or non-DOS) systems. Codegen86 (talk) 13:42, 15 April 2012 (UTC)
 * I've thought of adding limited, major parallel Apple (Motorola-processors) developments, but think the focus should be kept on x86 and x86 should remain in the title. There is already some mention of UNIX here.  Apple, UNIX and other OS could have their own timelines. —Wbm1058 (talk) 15:48, 15 May 2012 (UTC)


 * On second thought, you're right about x86. x86 DOS is the primary topic for DOS, so x86 is redundant. Per WP:PRECISION, concise titles are preferred, and over-precision should be avoided.  I'm moving the two violators of this policy, Comparison of x86 DOS operating systems and Timeline of x86 DOS operating systems—Wbm1058 (talk) 14:38, 15 June 2012 (UTC)


 * Geesh. Taking x86 out of the title to make it more concise opened the door to the hatnote:
 * This article is about PC DOS/MS-DOS and compatibles. For the IBM mainframe operating system, see DOS/360 and successors.
 * I'm thinking that losing the x86 may not have been such a good idea. If one views hatnotes as part of the title in a way, then now the title is less concise with the hatnote modifier. "PC DOS/MS-DOS and compatibles" is a long winded way of saying "x86". Wbm1058 (talk) 14:23, 27 July 2013 (UTC)


 * List of DOS commands has similar titling issues. I moved the title back from List of MS-DOS commands since not all DOS commands are MS-DOS commands—some are unique to PC DOS. Linking x86 in the lead sentence and adding an explanatory message to the article headers, "Other DOS operating systems are generally not part of the scope of this..." seems to solve the issue. Wbm1058 (talk) 18:27, 4 November 2013 (UTC)

DOS 3.1
What is the source for MS-DOS 3.1 being released in November 1984? Is it the InfoWorld article from Dec 3, 1984? That one only says that MS-DOS 3.1 will be appearing "within the next few months". Then it talks about something having been released by AT&T in Nov '84 in Europe, but it's rather unclear whether that was 3.0 or 3.1 or something else (3.05? 3.06?). Since IBM's DOS 3.1 didn't show up until April 1985, it would be good to have some better source than a poorly written InfoWorld article which can't get the versions straight to begin with... --88.67.95.53 (talk) 08:16, 7 May 2012 (UTC)
 * On page 39 of The MS-DOS Encyclopedia: "...so in August 1984, Microsoft released version 3.0 to IBM without network software."
 * On page 43, "...Version 3.1, completed by Zbikowski and Reynolds and released three months later, completed this network support...".
 * Three months later is November 1984. I'll update the article to add the reference. – Wbm1058 (talk) 15:34, 15 May 2012 (UTC)
 * Keep in mind that IBM's network software differed from Microsoft's, as it supported proprietary IBM hardware. That seems to me the main reason for the delay in the release of IBM's version 3.1 – Wbm1058 (talk) 15:48, 15 May 2012 (UTC)
 * I found the MS-DOS Encyclopedia entry in the meantime... it's poorly edited and could be read as "Zbikowski and Reynolds completed version 3.1, which was then released 3 months later". You should perhaps quote Advanced MS-DOS Programming instead which says: "MS-DOS version 3.1, which was released in November 1984...". The bigger question is what "released" means here. Released to OEMs? There were no end-user MS-DOS versions back then. When did OEMs actually ship it? If you say that MS-DOS 3.1 was released in Nov '84 but no customer could actually buy it, isn't that misleading? Consider also http://actapricot.org/disks/aprid5ks.htm#apr00029.dsk which contains MSDOS.SYS dated 11/28/1984 and which is MS-DOS 3.06, not 3.1, yet supports MS-NET 1.0. Speaking of which, what's the source for MS-DOS 3.05 being released in August 1984? Codegen86 (talk) 11:18, 19 May 2012 (UTC)
 * I removed mention of 3.05 with this edit, since it's just a minor version number and no specific date could be confirmed. The only (reliable) source I've seen mentioning that version number is Duncan's encyclopedia, p. 39: "These features were made available to Microsoft's other OEM customers as version 3.05." I trust that the article is not misleading about MS-DOS versions before 3.2 as the article lead clearly states that these were just OEM adaption kits. I think it's generally beyond the scope of this timeline to include version numbers of specific OEM adaptions, beyond the major ones like IBM and Compaq. The difference between 3.0 and 3.1 is a little muddy because, as sources clearly indicate, 3.0 was supposed to include networking, but, that wasn't ready yet when IBM shipped the AT. So, we have confused articles like this one, which can't quite decide whether to call Microsoft's November release 3.0 or 3.1, saying that DOS 3.1 is "essentially the same" as DOS 3.0 and "PC-DOS 3.0 also supports networks" (well yes, but it's only partial support that wasn't finished). One might thus assume that Apricot's 3.06 was just the network update to 3.05 (I see both of those version numbers on the page that you linked to) and that Apricot's OEM 3.06 is equivalent to IBM's OEM 3.10—which IBM didn't release until later, when their network hardware was ready to ship with it. Just an extension of IBM vs. non-IBM version number differences analogous to 1.1 vs. 1.25 – Thanks for your input. Wbm1058 (talk) 17:57, 4 November 2013 (UTC)
 * Also note that IBM announced both PC-DOS 3.0 and 3.1 on the same day, 1984-08-14, blurring the distinction between them. Wbm1058 (talk) 18:07, 4 November 2013 (UTC)

Sorry that I took so long to research this. I indeed found that MS-DOS 3.1 did not ship in the US until May 1985. I've updated the timeline to clarify this. It seems that MSFT never shipped a version to other OEMs before IBM released theirs; the only MS-DOS versions I've seen that were first-releases were those in Europe or Asia, where IBM did not compete; i.e., versions that IBM didn't want. Would be nice to find a source confirming that this was actual Microsoft policy; albeit a policy that neither party discussed publicly apparently. Wbm1058 (talk) 22:08, 29 August 2014 (UTC)

DOS 3.2
Interesting to see how the date(s) for versions 3.2 have evolved here: – Wbm1058 (talk) 20:11, 4 November 2013 (UTC)
 * IBM PC-DOS 3.2 required 128KB RAM, fit on one 720KB DSDD diskette, had file dates 12/30/1985. COMMAND.COM is 23,791 bytes and the box says the operating system was sold "As-Is". Maximum disk partition native to OS: 32MB. (IBM PC DOS [//en.wikipedia.org/w/index.php?title=IBM_PC_DOS&diff=63998128&oldid=63735800 15 July 2006 diff])
 * Clearly the editor looked up the file creation date of COMMAND.COM on the disk itself
 * PC-DOS 3.2 - 1985 - required 128 kB RAM and came on one 720 kB disk (IBM PC DOS [//en.wikipedia.org/w/index.php?title=IBM_PC_DOS&diff=93627630&oldid=89384816 11 December 2006 diff])
 * Reduction of detail from "file dates 12/30/1985" to just "1985" – now it's not clear where that date came from
 * PC-DOS 3.2 - December 1985 - required 128 kB RAM and came on one 720 kB disk or two 5.25 disks.  has 23791 bytes. (IBM PC DOS [//en.wikipedia.org/w/index.php?title=IBM_PC_DOS&diff=100359381&oldid=100186969 13 January 2007 diff])
 * Restored the month
 * Merged to Comparison of x86 DOS operating systems ([//en.wikipedia.org/w/index.php?title=IBM_PC_DOS&diff=181988650&oldid=177712516 3 January 2008 diff])
 * December 1985 PC-DOS 3.2 is released. (Comparison of x86 DOS operating systems [//en.wikipedia.org/w/index.php?title=Comparison_of_DOS_operating_systems&diff=182016606&oldid=182014976 4 January 2008 diff])
 * Merged from IBM PC DOS (now the "morphing" of file date to release date is complete)
 * Moved timeline to: Timeline of x86 DOS operating systems ([//en.wikipedia.org/w/index.php?title=Comparison_of_DOS_operating_systems&diff=247800631&oldid=243660647 17:07, 26 October 2008 diff])
 * At 14:17, 13 September 2011, after finding a source for when PC DOS 3.2 was actually released, I removed the entry for December 1985 ([//en.wikipedia.org/w/index.php?title=Timeline_of_DOS_operating_systems&diff=450297697&oldid=450221652 diff])
 * IBM had two announcement dates for 3.2, the first when the Token Ring Network was announced, the second when the IBM PC Convertible was announced a little while later. They held off on announcing the 720 KB floppy disk capability, even though Microsoft added it months ago, until they could also announce hardware that used it. But the 12/30/1985 file dates reported by the first editor are consistent with a January Microsoft release.

OK, I've updated the details for the 3.2 release. I think we have a better understanding of the timing and reasons for it. With 3.0 the unfinished software was rushed out the door because IBM's hardware was ready and waiting for software. With 3.1 and 3.2, Microsoft's finished software had to wait on IBM's hardware and software before it was released to the public in the US. I decided this editorial wasn't appropriate for including in the timeline, so I'll just link to it here: [//books.google.com/books?id=nvgwHu6zOfQC&pg=PA16 When confusion reigns supreme.] Microsoft's spokesman didn't even know that IBM had released DOS 3.2 and refused to talk about the product. Confirmation of Microsoft waiting for IBM to go first, before announcing general releases for other OEMs. Wbm1058 (talk) 12:35, 30 August 2014 (UTC)

Notability of PTS-DOS
Is PTS-DOS really notable enough to be listed ? It is a little known product with only 3 entries in the timeline. Asmpgmr (talk) 16:30, 15 June 2012 (UTC)
 * There are a couple other lesser-known DOS operating systems listed, i.e., PC-MOS/386 and ROM-DOS. My opinion is that if they are notable enough to have their own Wikipedia articles, they are notable enough to be listed here.  There once was an "other" category in the color key (it was white), and its only member was ROM-DOS.  Since then the white background came to be used for notable DOS-era developments, such as expanded memory.  I've brought back the "other" category, and merged PTS-DOS into it. – Wbm1058 (talk) 19:10, 15 June 2012 (UTC)
 * Should FreeDOS also be put in the other category ? It has few entries in the timeline and it's definitely not as significant as DR-DOS which is the primary compatible DOS. Asmpgmr (talk) 19:31, 15 June 2012 (UTC)
 * I'm inclined to let FreeDOS be for now. It is about the only DOS still actively supported, it seems. Though the frequency of each release is more reflective of how easy it is to distribute a new version online, versus package and ship floppy disks. Wbm1058 (talk) 00:56, 20 June 2012 (UTC)
 * While PTS-DOS (and S/DOS, which was basically the same) left a lot to be desired in regard to compatibility with undocumented APIs, I absolutely think it was notable. Taking much freedom in the interpretation of how DOS should look like (internally and on user level) and adding a huge amount of pragmatism when it came to the actual implementation, they were able to put together a DOS clone with a very small memory footprint (even smaller than DR-DOS), high performance and configurability. It shrugged me, though, when I looked more closely - the code looked incomplete and unfinished and there were lots of really dirty hacks which I expect to have caused serious compatibility and maintenance issues. Nevertheless, they had numerous nice ideas for extensions. I don't know about the US, but for several years it had some following among DOS users over here in Western Europe, probably even more so in Eastern Europe. By far not as much as DR-DOS, but certainly enough to be notable, I think. --Matthiaspaul (talk) 16:03, 4 July 2012 (UTC)

Multitasking MS-DOS 4 and PC DOS 7.1
How much is known about these two versions of DOS ? Are there definitive dates for their release ? Is there any definitive technical info available for either ?

Multitasking MS-DOS 4 (aka European MS-DOS 4)
 * International Computers Limited (ICL) was an OEM
 * Apricot, Groupil and Siemens may have been OEMs as well
 * 4.0 was based upon DOS 2
 * 4.1 included changes from DOS 3.1
 * ultimately formed the basis of OS/2 1.0

References from Larry Osterman who actually worked on it:
 * http://blogs.msdn.com/b/larryosterman/archive/2004/03/22/94209.aspx
 * http://blogs.msdn.com/b/larryosterman/archive/2009/08/26/24-years-ago-today-1985.aspx

Ralf Brown's Interrupt List mentions some of the enhanced Int 21h API functions which supported multitasking.

PC DOS 7.1
 * never officially released
 * several revisions between 1999 and 2003 (last was build 1.32 from December 2003)
 * based upon PC DOS 2000
 * major new feature is FAT32 support
 * was used in several versions of Norton Ghost and IBM's ServerGuide Scripting Toolkit

Asmpgmr (talk) 21:45, 17 June 2012 (UTC)


 * There is very scant information on multitasking MS-DOS 4 (I'll call it M-DOS 4) because it was more or less an unreleased product. Larry Osterman is essentially the only source of information. One of the OEMs was SMT Goupil (not "Groupil"), part owned by France PTT. Apricot and ICL OEMed M-DOS 4 as well. According to Larry Osterman, Goupil was the only OEM who eventually received multitasking MS-DOS 4.0 and ran it on the Goupil G4. ICL DRS 300 was one of the multitasking MS-DOS 4.1 systems (Google for it). I don't think it's very accurate to say that M-DOS 4 "formed the basis of OS/2". Not much more than DOS 3 did. M-DOS 4 pioneered some technologies also used in OS/2, such as the New Executable format (also used by Windows), detached programs, or pop-ups. But M-DOS 4 did not support multi-threading, virtual memory, or a GUI, and most importantly did not run in protected mode. M-DOS 4 was more like a fork, as evidenced by the fact that ICL shipped multitasking MS-DOS 4.1 in September 1987 while OS/2 1.0 was released only about two months later. Codegen86 (talk) 14:07, 4 July 2012 (UTC)

21:08, 17 June 2012‎ edit which removed stuff about multitasking from EMS 3.2 description
Duncan, Ray (1988). The MS-DOS Encyclopedia, Microsoft Press. ISBN 1-55615-049-0, p. 305:
 * The EMS version 3.2, modified from 3.0 to add support for multitasking operating systems, was released shortly afterward as a joint effort of Lotus, Intel and Microsoft.

It's not completely clear to me whether this release was concurrent with MS-DOS 3.2—note the Nov. 1985 item: "Both the Lotus-Intel-Microsoft and AST expanded memory specifications enable users to address up to 8 MB of RAM."

Page 316 of the same Duncan book:
 * The EMM relies heavily on the good behavior of application software to avoid the corruption of expanded memory. If several applications that use expanded memory are running under a multitasking manager, such as Microsoft Windows, and one or more of those applications does not abide strictly by the EMM's conventions, the data stored in expanded memory can be corrupted.

Maybe Duncan should have said, "The EMS version 3.2, modified from 3.0 to add support for multitasking operating environments..."? What he said implies OS/2, what he should have said implies Windows, TopView, etc.? – Wbm1058 (talk) 04:32, 18 June 2012 (UTC)


 * Multitasking operating environment (meaning Windows 3.x) makes much more sense. OS/2 didn't use expanded memory. Microsoft used the terms "operating system" and "operating environment" interchangeably in describing Windows 3.x which is unfortunate. Also OS/2 1.0 was released in December 1987 though it was announced in April 1987 (with the first PS/2 models). Anyway since EMS was designed for DOS programs is there any reason why multitasking should be mentioned in the DOS timeline ? Asmpgmr (talk) 05:38, 18 June 2012 (UTC)


 * There are other mentions of multitasking in this timeline. Your thoughts on the title of this article (see above section)? Timeline of the DOS era? Efforts to market multitasking solutions, particularly at Digital Research, span almost the entire era.


 * The definition of "multitasking" in the context of DOS is very vague. It might include certain TSRs (PRINT.COM), and it certainly should include peer-to-peer network servers like the IBM Network Program. If one were to talk about multitasking in the DOS context, it would have to start(?) with PRINT.COM background spooling and center on DOS 3.0/3.1, critical sections within DOS, and the SDA (Swappable Data Area) in DOS 3+. The EMS function to save/restore page maps was most likely intended for Windows (Windows 2.0 used EMS a lot when it could), but also would have been very handy for any TSRs utilizing EMS. And yes, OS/2 1.x had no support for (or use of) EMS whatsoever. Codegen86 (talk) 14:27, 4 July 2012 (UTC)

--- Timeline of x86 multitasking (incomplete, need to add mention of earlier operating environments)
 * April 1982: Digital Research announces Concurrent CP/M-86, aka Concurrent CP/M, a new CP/M-86-compatible single-user multitasking operating system
 * October 1983: Digital Research releases CP/M-86 Plus Version 3.1, based on the multitasking Concurrent CP/M kernel
 * August 1984: IBM announces a new multitasking windowing software utility called TopView
 * January 1985: Digital Research previews Concurrent DOS 286, an 80286 native mode operating system, allowing users to take full advantage of the protected mode to perform multi-user, multitasking operations
 * August 1985: IBM and Microsoft announce a Joint Development Agreement to create a new multitasking operating system from scratch, known as Advanced DOS (I've also seen "286 DOS")
 * (so Duncan in his book may be incorrectly implying "Advanced/286 DOS", anyhow my copy of his book is copyright 1988)

---
 * 4th qtr. 1986: Microsoft releases MS-DOS 4.0, the first multitasking version, to European customers
 * April 1987: Foxbase 2.0 386 does not support multitasking, which must wait for advances in the operating system
 * May 1990: DOS Protected Mode Interface (DPMI) version 0.9 is formally released—unlike VCPI, DPMI was designed for a multitasking operating system (hmmm, do they mean an operating environment like Windows 3.0??)
 * A lot of timelines on Wikipedia give little more than version numbers and dates. This timeline attempts to explain details of the versions and give an overall context.  Maybe it can try to address misconceptions about what multitasking means.  Terminate and Stay Resident "creates the appearance of multitasking," so an end-user may believe they are multitasking when they call up Sidekick. (Don't get me started on whether a human can really multitask, and whether they are operating efficiently when they timeshare a dozen tasks with their Franklin Planners, vs. lock the door and keep complete focus on the programming task at hand ;) Wbm1058 (talk) 13:35, 18 June 2012 (UTC)


 * Concurrent DOS really was multitasking. TopView was an IBM program which attempted to provide this as well. Of course European MS-DOS 4.0 really did have multitasking support. Advanced 286 DOS was the early designation of what became OS/2. Those entries are fine. I think the Foxbase sentence regarding multitasking should be removed. As for DPMI, well it is the DOS Protected Mode Interface and was primarily designed to allow programs to be written which ran in protected mode and use extended memory while being able to utilitize the DOS API. Windows 3.x provided a DPMI server and Windows 3.x was multitasking so it would have to be referring to that. I think the "designed for a multitasking operating system" part should be removed.


 * A separate issue: what is the source of the April 1986 date for European MS-DOS 4.0 ? I know that is the correct timeframe but I could never find a specific date of its release. Asmpgmr (talk) 15:33, 18 June 2012 (UTC)
 * My source is the InfoWorld article referenced in the timeline. Search for it on Google books.  I'm afraid it doesn't give the definitive release date you're looking for.  In the March 24, 1986 issue, they said that 4.0 "will go on sale to European customers in April."  I suppose that leaves open the possibility of unexpected delays, and I'm making a definite statement in the timeline, but really, announce the product will be out in April a week or so before April, and then miss by over a month?  Microsoft acknowledged the existence of 4.0, but said it is an unannounced product that had been preannounced by Apricot. Apricot said they had worked with Microsoft for nearly two years developing the Apricot implementation of DOS 4.0—Wbm1058 (talk) 23:34, 18 June 2012 (UTC)
 * Well I just searched some more and found an article that seems to make "will go on sale to European customers in April" seem like a fabrication or completely inept misstatement. I'll update the timeline to reflect this new information—Wbm1058 (talk) 00:01, 19 June 2012 (UTC)
 * Ralf Brown's Interrupt List said that it came out between DOS 3.2 and DOS 3.3 so if that info is correct that puts it between April 1986 and April 1987. I've read that it came out sometime in 1986 but I don't have a definitive source. Supposedly Multitasking MS-DOS 4.1 came out in 1987 but again no definitive source. Asmpgmr (talk) 00:36, 19 June 2012 (UTC)
 * We have the European 4.0 release narrowed down to the last 3 months of 1986. The InfoWorld Mar 2, 1987 issue has a table on page 5 that confirms: "DOS 4.0 Released in Europe in 1986"... haven't found anything yet that gets more specific than that 3-mo. window.  Did find someone posted a picture of the Oct 1987 4.1 diskettes and listed their contents. Nothing similar for 4.0 on that site, their 4.0 is IBM's version—Wbm1058 (talk) 00:56, 20 June 2012 (UTC)


 * From Osterman's Bio, it seems the 4.0 project was mostly complete by March 1986, probably about the time that Microsoft pitched it to IBM and Apricot pre-announced, expecting it to be finished soon. But then IBM wasn't interested, and Apricot, being a smaller customer, had to wait "about 6 months" while Osterman's skeleton crew "finish(ed) up DOS 4 - mostly spent working on bug fixes and cleaning up the remaining features that had to be completed." –Wbm1058 (talk) 01:31, 20 June 2012 (UTC)


 * Is this sentence really necessary: Application programs communicate directly with the EMM using software interrupt 67H, bypassing DOS. New INT 67H function 4EH supported multitasking operating systems by saving and restoring page maps
 * It is a bit technical to be in a timeline and inconsistent since there is no technical info regarding changes to the DOS Int 21h API listed anywhere in the timeline. Asmpgmr (talk) 17:21, 19 June 2012 (UTC)
 * OK, I edited it to remove the hex numbers. I probably went a bit overboard because the issue of whether the EMM supported a multitasking O/S was raised.  Also I was forced to tie the EMS 3.2 release to DOS 3.2, which was misleading since EMS happened months earlier.  Could never find an article about MSFT coming onboard in InfoWorld.  But now Google has more magazines searchable online, so I finally found a source for the timing of LI becoming LIM—Wbm1058 (talk) 00:56, 20 June 2012 (UTC)
 * I thought that EMS 3.2 increased the memory limit to 8 MB from 4 MB in EMS 3.0. Not 100% sure since this was a long time ago. I know that EMS 4.0 increased the memory limit to 32 MB and this can be verified since the EMS 4.0 spec can still be found online. Note I'm pretty sure there was only about half a year between EMS 3.0 and EMS 3.2 Asmpgmr (talk) 02:57, 20 June 2012 (UTC)


 * Found this interesting link - http://ps-2.kev009.com/bocaresearch/Drivers&Software/Archived/File_area_20/BRATPLUS.TXT Asmpgmr (talk) 03:19, 20 June 2012 (UTC)

VESA BIOS Extension added to timeline
I added VBE 1.0 to the timeline. I can't find the original spec but here is a site that has most of the VBE specs from 1.0 to 3.0 - http://nightmist.us/study/SL/SUITE/VBE/ Asmpgmr (talk) 04:47, 20 June 2012 (UTC)
 * VBE 1.0 - October 1989
 * VBE 1.1 - ? 1990
 * VBE 1.2 - October 1991
 * VBE 2.0 - November 1994
 * VBE 3.0 - September 1998


 * Unfortunately the nightmist.us link appears to be broken. Anyway, http://ps-2.kev009.com/bocaresearch/pdf/manuals/SVGA_ALL.PDF suggests that the VBE 1.1 spec was published in June 1990 (#VS900602, cf. RBIL). Codegen86 (talk) 12:32, 5 July 2012 (UTC)

Out of control
This article appears to be getting out of control and morphing into something that's definitely not a timeline. A prime example is the August 1981 entry -- information like the use of "Mostek MK4116 16-kilobit 16-pin DIP chips" in the original PC is irrelevant here; such trivia should perhaps be included in the IBM PC article but not in a timeline that aspires to cover a span of several decades and a multitude of software and hardware products. Codegen86 (talk) 12:10, 9 July 2013 (UTC)
 * I agree with you that it's a bit of a borderline case, however, in general I very much appreciate the article's primary contributor's approach to enrich this timeline with many bits of information, which help to put the various developments into a broader context. The result is more than a simple list of DOS versions, it is a condensed, but still reasonably good readable historical abstract over a certain period of time with the focus on DOS and what people, who used it, had to deal with. Including some "nostalgia" information beyond the core facts is just what makes or keeps this article interesting to read. So, while mentioning the RAM vendor here may be a bit too much, mentioning, f.e., the exact model number of the floppy drives used, is very useful information. I do not agree with your "out of control" classification, though. --Matthiaspaul (talk) 13:34, 9 July 2013 (UTC)
 * The information is certainly interesting and sometimes useful, but the question is whether this is the best place for it. It often makes the timeline harder to use as a timeline. The idea of hypertext is that information can be linked rather than just bunched together, and this is a great example of a situation where linking should be used. That way, the reader can choose whether to learn about the minutiae of a specific product but the actual timeline is not obscured by too much detail. Just loading and formatting the article now takes significant time, which is probably a good sign that there's too much. Just saying. Codegen86 (talk) 10:24, 10 July 2013 (UTC)
 * I see your point and even agree with it to some extent, I just don't think anything is "out of control" here. That's much too negative (to the point it may personally hurt the main contributor), while I consider the existance of this article (in the current form) as a big plus for Wikipedia. Someone really devoted time and energy to go through all the sources. As such it has meanwhile become a very good quick reference also for other article development.
 * We would certainly not duplicate the level of detail in other lists, but since we already have various other lists for DOS versions alone (f.e. in the MS-DOS and PC DOS articles, although they are incomplete, badly sourced and sometimes inaccurate at present), I don't think we should strip this list down into yet another one of those, but rather complete/correct the other lists with information provided here. --Matthiaspaul (talk) 13:31, 10 July 2013 (UTC)

PC-DOS
I appreciate your support in the above section, and intend to respond there later. But first, I want to discuss your recent edits. "PC-DOS was a trademark of IBM's rival DEC. Likewise, DRI also used PC DOS rather than PC-DOS" I'll propably have more comments later. Cheers, Wbm1058 (talk) 16:26, 29 July 2013 (UTC)
 * PC-DOS is not a "spelling mistake". It is a style decision. There is no right or wrong answer here, just a consensus on what works best. We tend to favor common names, as used by a significant majority of reliable English language sources, over official names. Consistent style would use either MS-DOS and PC-DOS, or MS DOS and PC DOS. But sometimes other considerations override style.
 * What is your source for "PC-DOS was a trademark of DEC"? There is no mention of a DEC "PC-DOS" in Digital Equipment Corporation, the only DEC-proprietary DOS mentioned there is DOS-11. Both live and dead trademarks may be confirmed by researching the primary source data at the United States Patent and Trademark Office Trademark Electronic Search System (TESS). Click on "Basic Word Mark Search (New User)". A search on "PC DOS" finds only one related trademark, the abandoned "CONCURRENT PC-DOS". So, note that Digital Research trademarked their version using a hyphen, perhaps in recognition of the common usage by the trade press.
 * It's also interesting to note the trademark registration for MS-DOS. " FIRST USE: 19800600. FIRST USE IN COMMERCE: 19811200" ... in other words, Microsoft did not use the term MS-DOS commercially until December 1981, after the release of the IBM PC. They just began using the term internally in June 1980 (I wonder what the context of that use was, the month before IBM first contacted them). And they didn't actually file for the trademark until March 15, 1984.
 * Look closely at the January 6, 1981 license agreement, and note that it is not specifically a license for version 0.3 as your modified footnote description implies. It is simply a license for 86-DOS, as documented in the documentation for version 0.3 – the documentation is dated 11-15-80, which says nothing about when version 0.3 was released (maybe before it was documented) and is not a significant date relative to the negotiation of the agreement. And version 0.3 is already old at the time of the license agreement, which includes the modifications to verison 0.3 as listed in Appendix "A" of the agreement. See how informal it all was, they agreed to license a product for which the formal documentation had not yet been updated!
 * Hi Wbm. regarding the DEC trademark, this was related to me by someone who should know the truth as the reason why IBM named the operating system "IBM Personal Computer DOS", originally. At some later stage, either the DEC trademark expired or they negotiated with DEC over this, as IBM later switched to use the term "PC DOS", officially.
 * I will check some old books for this as well, but it will take time. A quick web search revealed a copy of Dave Williams' "Programmer's Technical Reference", which supports my statement as well. I seem to remember, that an old version of RBIL had a similar statement, but I could not find this in the lastest version (and since I remember that I cleaned up the trademark section in there in one of the 50s revisions to reflect the then-current state of affairs rather than historic information, I may be even responsible for removing it there. I will have to look up older revisions in my archive.) (There are a number of other online references stating this as well (like this one ), but they seem to be either based on Williams' work or generally contain much less accuruate information and are therefore considered less reliable by me.)
 * I won't put too much weight into online databases for data this old. While a great research tool in general (for data in there) I have found that such databases are not necessarily complete. Also, since both DEC and IBM are international companies, we don't know, if this was an issue in the US or in another country. Nevertheless, I would like to find a definite proof as much as you do.
 * Regarding DRI's "Concurrent PC DOS", I know for sure that Digital Research deliberately did not use a hyphen in names such as "Concurrent PC DOS" and "DR DOS" (nevertheless you will find a few exceptions), a tradition held to until the name change from "OpenDOS" to "DR-OpenDOS" and then to "DR-DOS" in 1997-12 by Caldera. The reason for this name change to the (much unloved) spelling with hyphen was that it was legally difficult to register the trademark without the hyphen for a company named completely different, that is, what still might have worked for Digital Research (DR), did not work for Caldera. I don't know, but the fact that you found the term "Concurrent PC-DOS" registered (and not "Concurrent PC DOS") might have had similar reasons (although, as explained above, it does not necessarily mean "Concurrent PC DOS" was not registered as well), nevertheless, the official spelling is "Concurrent PC DOS".
 * Regarding Microsoft licensing 86-DOS, yes, I am aware of the fact, that they licensed all versions, not a particular one. The point I was trying to make was not that they licensed a specific version, but simply to explain the exact dependencies between the 86-DOS, MS-DOS and PC DOS version numbers. Some literature lumps MS-DOS and PC DOS version numbers together incorrectly, which has caused much confusion. MS-DOS and PC DOS kept (mostly) synchronized version numbers between 2.x and 5.x, but this was not the case for the 1.x versions and also not for versions since 6.0.
 * In reality, there never was a MS-DOS 1.0, as the product was still named 86-DOS at this time (and owned by Seattle Computer Products, not Microsoft). The first (internal) version of MS-DOS was either MS-DOS 1.10 or 1.11, based on 86-DOS 1.10. And PC DOS 1.0 was not based on a non-existant "MS-DOS 1.0", but on either (internal) MS-DOS 1.10, 1.11 or 1.14 (see my former detailed comment on this), whereas PC DOS 1.1 was based on (internal) MS-DOS 1.24. The first version of MS-DOS licensed to other OEMs than IBM was MS-DOS 1.25. Besides other OEM versions, there was a version of SCP MS-DOS 1.25 for the SCP Gazelle, which was a continuation of SCP 86-DOS (based on the fact that it still supported the 16-byte directory entries in addition to the 32-byte directory entries introduced with 86-DOS 0.42: I don't know if PC DOS 1.0 still supported 16-byte directory entries as well, but if so, it was never documented; PC DOS 1.1 did not support them for sure), so even SCP no longer named their former product 86-DOS.
 * --Matthiaspaul (talk) 13:58, 4 August 2013 (UTC)
 * It doesn't appear that DEC ever registered (®) any "DOS" trademark. It's possible they used the trademark symbol, which is all Seattle Computer ever did with 86-DOS. I don't think they ever used "DOS" for anything proprietary that related to personal computers. I suspect they just may have used DOS-11™ back around 1970. For that matter IBM may have trademarked DOS/360. What exactly would this DEC "PC DOS" product have been? – Wbm1058 (talk) 13:56, 18 August 2013 (UTC)


 * Regarding your doubt about the handwritten date of the 86-DOS sale agreement, see 7 (number). Americans generally don't put a line through the middle of the number seven, or write a one with such a long tail so as to make it look like a seven without a line through the middle. – Wbm1058 (talk) 14:11, 18 August 2013 (UTC)

The first release of QDOS
Sources on the origins of QDOS relating to specific dates and versions are somewhat muddy. Hard Drive by Wallace & Erickson calls the product 86-QDOS while other sources simply call it QDOS. Hard Drive says "in September... Paterson had his operating system up and running for the first time." (p.183) Softalk March 1983 issue says "By July, he had finished roughly 50 percent of the operating system and called it QDOS 0.10 (for quick and dirty). He quickly found a bug and it became QDOS 0.11... By the end of August 1980, QDOS 0.11 worked well and was being shipped. It didn't stay QDOS very long, and not many copies were distributed." Encyclopedia of Computers and Computer History by Raul Rojas says "QDOS... began shipping with the 8086 computer system in August. SCP approached Microsoft about adapting their software to run under DOS, who said it was possible - for a price." Paterson himself, in the June 1983 Byte says "The first versions of the operating system, called QDOS 0.10, were shipped in August 1980." Gates by Manes & Andrews doesn't give version numbers or release dates, but says "...in early August Paterson wrote Bob O'Rear about adapting Microsoft BASIC for the new operating system.... Rod Brock, then wrote Paul Allen to propose a cross-licensing arrangement... Result: When the IBM emmissaries at the August 28 meeting asked about 16-bit operating systems, Bill Gates hinted that he might just be able to find one." (p.158) Hard Drive says, "Once Paterson had 86-QDOS working, he contacted Paul Allen and asked him if Microsoft wanted to adapt any of its software for Seattle Computer's new operating system." (p. 185) Now if the first DOS manual was released "ca Oct 1980", then the QDOS that shipped in August or September would have shipped without a manual.

What I take away from this is that apparently the most notable thing that "shipped" in August 1980 was awareness that Seattle Computer had a rudimentary, preliminary, 50-percent finished, working 16-bit operating system—knowledge shipped to Microsoft. If it had shipped with any fanfare or advertising, you would think IBM would have learned of its existence too...

Well, looky here. Paterson may still have been calling his OS "QDOS" in August 1980, but Seattle Computer is already running a full-page ad in Byte calling it 86-DOS and offering it for $95. I guess IBM's people never noticed that ad, or, if they did see it, didn't realize its significance as Gates & Allen did. Of course, they had the benefit of direct contacts from Paterson and Brock, perhaps before the ad was published, to help them realize its significance. – Wbm1058 (talk) 13:56, 18 August 2013 (UTC)


 * Regarding "Wallace & Erickson calls the product 86-QDOS", clearly they mixed up the name. It was originally internally "(SCP) QDOS", and later officially "(SCP) 86-DOS". Informally, it was also named SCP DOS. (And yet much later it was SCP MS-DOS, when they backlicensed it from Microsoft as an OEM.)


 * Your published "BYTE August 1980" source for "86-DOS" is quite interesting (Do they state the exact publishing date somewhere, possibly even end of July? And how many weeks overhead did they had for ads?), as it partially contradicts other sources putting the name change somewhen between QDOS 0.2 (1980-08-??) and 86-DOS 0.33 (1980-12-??). Softalk states it happened before 86-DOS 0.33 in 1980-12, other sources, however, already mention a 86-DOS 0.3 (1980-11-15). Knowing that 86-DOS 0.34 was released on 1980-12-29, we could narrow the name change down to either version 0.2, 0.3 or 0.33 and a possible date range between 1980-08-02 and 1980-12-28. The BYTE ad helps to shrink this to between 1980-08-02 and 1980-08-31. So, if 0.2 was not already named 86-DOS, 0.3 definitely was. (ToCheck: Including possibly confused sources putting 0.10 into August not July, even 0.10 and 0.11 might already have been named 86-DOS officially.)


 * Various documents state that 0.3, 0.33 and/or 0.34 were already given/licensed to Microsoft (probably for evaluation), but Microsoft definitely was not the only customer at this time. I don't think anyone received anything before version 0.11 (1980-08-??) or 0.2 (1980-08-??) (Paterson states, 0.10 was the first published version, putting it into August, although other sources indicate July). I'm not sure, if Microsoft got anything before 0.3 (1980-11-15) - perhaps for evaluation.


 * I very much assume, IBM was (made) aware of the existance of SCP's 86-DOS, but they wanted to deal with Microsoft rather than with SCP, given that SCP was a "nobody" and they already dealt with Microsoft in regard to the languages (and DRI was out of the deal). However, it was SCP, who were not aware of the fact that Microsoft wanted their OS for their deal with IBM (and thereby didn't realized its real value until much later). --Matthiaspaul (talk) 17:12, 18 August 2013 (UTC)


 * I'm also interested in learning the actual publishing date and ad deadline for the Byte magazine issue. I suspect that Paterson had a version good enough to prompt Brock to run the ad sometime in July, and that the ad was submitted to the magazine in July. Possibly the first publicly released version announced itself in its banner as QDOS 0.1 when it started up, but could have been labeled on the 8" floopy disk as 86-DOS. Seems Brock never liked "quick & dirty" and quickly changed the name, I'm speculating.


 * I'm not particularly concerned with trying to sort out a comprehensive list of minor version numbers, and it seems this is very difficult to reliably source. My opinion is that this site has it about right. There were just three major versions: 0.1, 0.3 and 1.0. There probably was an original 0.1 in July that was good enough for Brock to start the ball rolling with marketing, then a bug fix after the Byte ad ran resulting in the first version shipping being 0.11 – but they could have still called it 0.1 in practice, nothing worth reprinting diskette labels over. And after 1.0 shipped and Paterson started working at Microsoft, anything after that was likely minor bug fixes done on contract when he was "off-the-clock" at Microsoft, so to speak. Again, these minor details may be difficult to reliably source. Wbm1058 (talk) 17:55, 18 August 2013 (UTC)

Garish color coding
The colors chosen are way too saturated, lending a carnival poster appearance to the entire article. Still worse, the intense colors seriously impede readability for normal color-vision readers, and may be even more of an impediment to color-blind readers. Could somebody who understands basic graphic design choose some appropriate pastel colors for less disruptive color coding? Reify-tech (talk) 15:02, 17 August 2013 (UTC)
 * ✅ – I hear you. I just put up a version with brighter, more pastel colors. Hope you think it's better. Of course, white is the ultimate in brightness, and the more pastel the colors get, the closer to white. – Wbm1058 (talk) 19:40, 21 August 2013 (UTC)

Free download
At the moment the article states that FreeDOS was the first free download, or similar. Why isn't that Caldera OpenDOS? I got the idea that this was far from "freeware" as we understand it today, but published sources could be downloaded, couldn't they? –Be..anyone (talk) 17:49, 13 February 2014 (UTC)
 * Good catch! I fixed the lead section to mention both. Wbm1058 (talk) 01:31, 31 August 2014 (UTC)

Pageviews
I just found WikiProject Computing/Popular pages, which was last updated on April 14, 2016. To make the top 1,000 required an average of 832 views per day, and there are no timelines in that list. However, interest in this article seems to have spiked upwards lately (it's unclear to me why), which is great. Recently the views are averaging over 2,000 per day so if traffic on computing articles in general has been otherwise steady, then this article would be in the top 250 now!

Noting from the 2016 Community Wishlist Survey results that the wish to Fix and improve Mr.Z-bot's popular pages report ranked #9, so looking forward to the resurrection of that report so we can get an update on how "my page" is doing! wbm1058 (talk) 16:48, 6 January 2017 (UTC)
 * I just added a graph of daily page view counts at the top of the page, below the banners. Views have been steady in the ~500 range for most of the past year, with the exception of a 3 – 4 month spike centered around the beginning of 2017, when the views were at 3,000 or more for much of the time, for some unknown reason. Most curious. – wbm1058 (talk) 13:25, 22 May 2017 (UTC)


 * # 527 of 1,000 on the Wikipedia:WikiProject Computing/Popular pages report for October 2017! Yeah! wbm1058 (talk) 17:21, 29 October 2017 (UTC)

CP/M was first written in 1974, not 1973
Further research has confirmed that I was misled by the MS-DOS Encyclopedia about the date that Gary Kildall first wrote CP/M.

The CP/M article has it right:


 * Gary Kildall originally developed CP/M during 1974, as an operating system to run on an Intel Intellec-8 development system, equipped with a Shugart Associates 8-inch floppy disk drive interfaced via a custom floppy disk controller. It was written in Kildall's own PL/M (Programming Language for Microcomputers). Various aspects of CP/M were influenced by the TOPS-10 operating system of the DECsystem-10 mainframe computer, which Kildall had used as a development environment.

Whereas page 10 of the MS-DOS Encyclopedia says: "In 1973, Kildall... was working with an 8080-based small computer given him by Intel Corporation..." Not likely, as the Intel 8080 wasn't released until April 1974, and as Kildall was bartering his software for Intel hardware he probably didn't have special access to prototypes of the 8080 prior to its release, or he would have said so in his autobiography. Also, "The version of CP/M developed by Kildall in 1973..." No. He developed PL/M in 1973; CP/M came about a year later.

The MS-DOS Encyclopedia is generally a reliable source for things relating to Microsoft; it's less reliable for things relating to Kildall and CP/M. I'll update the article with more details from better sources. – wbm1058 (talk) 17:54, 29 October 2017 (UTC)

"Microcomputer" vs. "Embedded system"
In the early days of microcomputers, most were embedded systems. This Ngram shows that the term "embedded system" didn't start taking off as a common term until the 1980s, and this Ngram shows that "microcomputer" has always been more commonly used. "Microcomputer" began taking off coincident to Intel's early 1970s introduction of the microprocessor, and peaked in 1984–86. Only after home computers took off ca. 1977 did the primary meaning of "microcomputer" transform from "embedded system" to "home or personal computer". Before there were home and personal computers, there were just embedded and development systems, for the most part. Really, "microcomputer" should be thought of as a catch-all term that covers all of these. "Development systems" such as the Intellec were the precursors of the later personal computers. Like many home computers, they also had resident monitors rather than complete operating systems. – wbm1058 (talk) 15:48, 31 October 2017 (UTC)

Design of the CP/M file structure and command processor
Much has been said about how Tim Paterson "copied" or "cloned" CP/M to create DOS. Certainly it's something like that on the basis of the look and feel of DOS 1.0. DOS emulated the API of CP/M – that's a fair way to put it. So to say he "invented" DOS or is the "father of DOS" is a bit over the top. It's not that original, but it is a significant "evolutionary adaption", or "development", or "improvement". Very significant. The more I study the history of technology, the more I find that this, not truly original invention, is the norm. Technology evolves.

Note that per, Kildall based much of the design of CP/M's file structure and command processor on operating systems from Digital Equipment, such as RSTS/E for the PDP-11.

And if I looked hard enough, I wouldn't be surprised if I found that DEC developed that stuff based on even earlier developments in computing. – wbm1058 (talk) 20:03, 8 November 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Timeline of DOS operating systems. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20160311025324/https://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-53564 to http://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-53564

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 13:22, 20 November 2017 (UTC)

If everyone reading this page in the past month gave me $3
...I would have $78,000. Just sayin' wbm1058 (talk) 19:59, 9 August 2018 (UTC)

PC DOS 7.0 release date
I can't find a reference for this but I suspect the release date for PC DOS 7.0 was around November or December 1994, not April 1995. Time timestamp of the files on the installation diskettes is 1994-11-17. ozzmosis (talk) 08:20, 4 March 2023 (UTC)
 * You're right, it was earlier. Searching IBM's product announcement, I found: "Availability Date: February 28, 1995", which is the same date as the product announcement. InfoWorld reviewed the release in their April 10, 1995 issue, which is where I got the April 1995 date from. I'll correct this item in the timeline. Thanks, wbm1058 (talk) 22:09, 18 November 2023 (UTC)