Fantasy History at Ars Technica

Ars Technica today published an article titled “Half an operating system: The triumph and tragedy of OS/2“. Although very interesting, unfortunately the article to a significant extent engages in what can best be called fantasy history, which causes the text to contradict itself. And while it attempts to present a coherent view of the history of OS/2, the article is more a hodgepodge of facts/products/trends (PS/2, NT, PowerPC) picked at random to support an argument.

It also presents the old “geniuses at Microsoft vs. dumb IBM” story which is increasingly difficult to justify as time marches on. IBM is the bad guy, overly bureaucratic and incompetent. Yet listening to old IBMers a somewhat different picture emerges, with Microsoft being a disorganized, sloppy company producing poorly documented and buggy software. As always, there’s some truth to both sides of the story.

Let’s start with some of the inaccuracies. When OS/2 was conceived and designed, it wasn’t called OS/2 at all. In fact there is no known evidence that the name OS/2 existed before 1987. The name OS/2 had a lot to do with PS/2, but the OS itself not so much.

Then there’s this gem: “Today we take virtualization for granted […], but in 1985 the concept seemed like it was from the future.” Or at least it would have, if time flowed backwards and the future was 1972—see IBM’s own VM/370.

Then comes the the fantasy history. The 286 sucked—okay, that’s not an unreasonable argument. Therefore, OS/2 should have been designed to run on a 386. Now, hold on there. First off, 386 PCs didn’t even exist before 1986, and only covered a tiny fraction of the PC market when OS/2 was released. OS/2 was meant to be a mass market product, not a niche operating system guaranteed to be ignored by major ISVs.

Then there’s the minor detail that it took Intel a while to build 386s that were capable of running 32-bit code in reality instead of just on paper (the famous double-sigma CPUs).

But the biggest argument against 386-only OS/2 is given in the article itself: Memory. If 16-bit OS/2 1.x was considered to be too memory hungry, exactly how was bigger and slower 32-bit OS/2 supposed to solve that problem?

Yes, it took Microsoft and IBM much longer to complete 32-bit OS/2 than originally planned, but given that even in 1992, OS/2 2.0 was often considered to be too memory hungry, it’s hard to see how releasing 32-bit OS/2 years later would have improved things.

One of the biggest issues with the article is the claim that good DOS and Windows application support prevented native OS/2 applications from being written. That’s probably why Windows 95 didn’t support DOS and why Windows NT didn’t bother with a 16-bit Windows subsystem… right? The truth is that good DOS and Windows support was a necessary evil and without it, OS/2 would have fared just like the other PC operating systems with no or minimal backwards compatibility. NeXTSTEP, Solaris, UnixWare, and the like.

The article several times falls into the trap of assuming that people working for the same company must necessarily share the same goals. That wasn’t true of Microsoft (Windows 9x vs. NT), and it definitely wasn’t true of IBM, where several hardware groups fought each other and didn’t necessarily see eye to eye with software divisions. Once Microsoft quit OS/2, IBM was in an unfamiliar and difficult position of having to support competitors’ hardware, and competing against Microsoft on IBM’s own hardware. The inevitable internal infighting didn’t help OS/2, but that doesn’t mean IBM as a whole behaved irrationally.

Microsoft’s anti-competitive tactics are not mentioned, despite being very well documented. Arguably once Microsoft got a stranglehold on OEMs and forced them to preload DOS and Windows (securing a guaranteed source of cash flow), the game was over and no one was going to change that, not IBM and not anyone else.

Windows NT is presented as some kind of mythical software that magically ships a bug-free version 1.0. Most people don’t remember NT 3.1 as “multiplatform, ridiculously stable and fault-tolerant”, but rather as bloated, slow no-show that hardly anyone ever ran (NT 4.0 was certainly a different story!). And calling the PowerPC version of NT a “consumer operating system” seems to stretch reality more than a little.

Looking back from 2013, the answer to the question of which company was smart and which was stupid is less and less clear. IBM cut its losses with OS/2 and later on, sold its PC Company while it was still worth something, rather than waiting like HP and Dell. The article says that “IBM threw in the towel on OS/2 and then on PCs in general” as if that were a dumb thing to do… but the reality is that getting out of the PC hardware market when IBM did was was a very smart move, and killing OS/2 was more or less the only option IBM had, especially after the spectacular failure of Workplace OS.

IBM may be 100+ year old, but you don’t get to be that old without having learned a trick or two. Microsoft bet the farm on Windows, but Windows is not sexy anymore and making money with operating systems is increasingly difficult. In another decade, we might more clearly see who was really smart.

This entry was posted in IBM, Microsoft, OS/2. Bookmark the permalink.

50 Responses to Fantasy History at Ars Technica

  1. Thanks for this Michal, good points. As usual there is not one single view on the whole history. I was thinking about giving examples how OS/2 was used back then, I’ve seen so many strange places where OS/2 was used, sometimes for more than a decade on the exact same hardware in mission-critical environments and never failing (until the hardware started falling apart). Also what we probably did first with the WPS was to change it to strange colors and fonts but we all know that this was just scratching the surface. XWorkplace and other extensions allowed things I haven’t seen ever since. I think I still complain about other desktops environments (MacOS, Gnome, KDE…) at least once a month 🙂

    Was nice to read something from you, has been a while since we saw each other 🙂



  2. Hi Michal!

    Thanks for your comments and criticisms. I hope you don’t mind if I criticize some of them. 🙂

    It should be noted that I never once used the word “dumb” to describe IBM. If that image is created of the company in terms of its relationship to Microsoft, I would argue that that image (and that word) arrived in the mind of the reader, not the author.

    I’m not sure what you are criticizing when you are talking about my short discussion about the name OS/2. It clearly was chosen to go alone with the PS/2. That’s all I was claiming, and in your last sentence you agree with me, so I don’t see the problem there.

    The issue about virtualization is a misunderstanding, I think. I could have been more clear, and for that I apologize. What I was saying was that virtualization on the *PC* was a new and scary idea for IBM. The 386 was the first chip with these kinds of powers, and clearly showed that the way forward for the PC was to be able to perform mainframe-like virtualization tricks. I addressed this confusion in the comments, and I reserve the right to update the article itself to remove this confusion.

    I don’t think that writing 32-bit code was necessarily something that resulted in inevitable code bloat. After all, AmigaOS was 32-bit code internally and had preemptive multitasking and ran in 256KB of RAM. 🙂 Nevertheless, you make a good point that a 386-specific OS in 1987 would have had a small potential market. But the fact is that IBM didn’t get around to releasing a 386-capable OS/2 until 1992, by which time 386 computers were commonplace. If they had started earlier, they would have had such an operating system ready by this time. And the 286 was indeed brain-damaged.

    I disagree with your assessment of DOS/Windows compatibility. I am convinced (having lived through it at the time and discussed it with ISVs) that it doomed the market for native OS/2 applications. The real problem was that OS/2 applications offered very few advantages (long filenames, mainly) over Windows applications run under Win-OS2. They tended to be slower and buggier.

    You did make a good point about Windows NT also having backwards compatibility, but that was for an older version of Windows, with very easy ways for ISVs to update their apps to be 32-bit. (There was the whole Win32s API, for example, which would run on both Windows 3.1, Windows 95, and Windows NT, easing the transition). Still, at the end of the day, the argument about DOS/Windows compatibility is an opinion, and thus cannot be dismissed as “fantasy” as you have done in your criticism. We must simply agree to disagree. 🙂

    Getting back to the conclusion about IBM “throwing in the towel” on the PC market, this is also a matter of opinion. In terms of controlling the PC space, IBM competed directly with Microsoft for a few years and then gave up in complete and utter defeat. This isn’t a statement about “dumb” or “smart” but an analysis of IBM’s will to compete in this market. They clearly didn’t have it.

    In conclusion, I think that some of your criticisms are valid, but that to call my article “fantasy” is unreasonably negative, and I would disagree with this conclusion. Of course, I’m understandably biased. 🙂

  3. Andrea76 says:

    Compliments, very interesting article.

    I remember the slogan “Better than DOS, better than Windows and better than OS/2” (not sure it’s written right, do you that famous advertisement?)

  4. Michal Necasek says:

    Hi Adrian, nice to see you again 🙂 Yes, the WPS spoiled people, especially with extensions. And I’ve seen my share of OS/2 installs hidden in all sorts of custom machinery. Sometimes I wonder when the last non-virtual OS/2 system will die — it could be quite a while.

  5. Michal Necasek says:

    Hi Jeremy!

    Nice of you to stop by 🙂 The image of clever little Microsoft riding the stupid IBM bear headed for the cliff is a pretty clear impression I got from the article. Perhaps I was the only one, perhaps not.

    The name issue is that the text reads as though the OS/2 name was chosen first and then the OS designed. But the OS design goes much further back, and in some ways even predates the MS-IBM JDA (1985). The name “OS/2” was reportedly chosen very shortly before OS/2 was announced, and in fact some of the printed documentation that came with MS OS/2 SDKs (May-ish 1987) still talked about “DOS 5”.

    I honestly don’t know how scared IBM was of the virtualization capabilities of the 386. At the end of the day, the 386 could only virtualize an old 8086, not itself; nothing like what was proposed here. There’s good evidence that by 1991, IBM wholeheartedly embraced the idea. That’s all I know.

    AmigaOS was written for the 68000, and although it originally ran with a relatively tiny amount of RAM, it did have a rather big ROM (512K?). I suspect that neither IBM nor Microsoft ever seriously considered that possibility, for better or worse. On x86, 32-bit code is almost always bigger than 16-bit, and slower except in cases where large amounts of data need to be processed. It’s a simple consequence of the fact that the instruction encodings are longer and pointers are bigger. A similar effect occurs with 64-bit code, although there the speed issue (not code size!) is mitigated by twice the number of GPRs.

    I kind of agree with your point that good DOS/Windows compatibility hurt OS/2 in that fewer native apps got written. But I also believe that this was a) offset by the higher number of users who picked OS/2 because of that, and b) IBM simply had no choice. One of the weakest points of OS/2 1.x was poor DOS support, and it just needed to improve in 2.x. It’s definitely a dilemma because no compatibility means no users and too much compatibility means no native apps. I’m not sure optimal balance always exists.

    FWIW, I found OS/2 2.x/Warp to be an excellent development platform for DOS and 16-bit Windows apps. The real thing was just too fragile and too easy to take down by one runaway pointer.

    Win32s was clever for sure. Microsoft did learn a lot from OS/2, and they recognized that the lack of easy portability hurt OS/2 1.x. Of course they did work on things like WLO but all that came too late to make any difference. With NT, Microsoft was much better prepared, and for many years, MS could afford to sink millions if not billions into NT development without having to worry about their bottom line. And yet they didn’t think NT was good enough to put on all new systems until what, 2001?

    Don’t get me wrong, I hated IBM for “throwing in the towel” on OS/2. But in retrospect, the difference between IBM and Microsoft was that MS absolutely needed Windows to survive, while IBM did quite well without OS/2.

    One of Microsoft most brilliant long-term strategic decisions was not selling PCs; IBM was always going to compete with other OEMs on the hardware front, so why should those OEMs depend on IBM for software? Few people probably also appreciate how strong the anti-IBM and pro-MS sentiment was in the late 1980s and early 1990s, and just how strong Microsoft’s position was in the PC market (first riding on IBM’s back, then cajoling and coercing with no sense of ethics). At some point, the top IBM management sat down and came to the inevitable conclusion that fighting Microsoft was a losing proposition. I still don’t like that they killed OS/2, but I also recognize it was the smart decision. It’s not so much a question of will but seeing reality for what it is, I’d say. A will to fight a tsunami is surely a noble thing, but not a good survival trait.

    Basically knowing what I know now, I’m not at all convinced that there was a way for IBM and OS/2 to “win”. Anyway… if people still write about OS/2 so long after IBM decided to abandon it, perhaps IBM did something right after all 🙂

  6. dosfan says:

    The Ars article did come across like an indictment against OS/2 specifically and IBM in general while painting Microsoft as great. OS/2 did have its faults and so did IBM but Microsoft was also far from perfect. Several points about this article:

    * The title of article “Half an operating system” – supposedly because “OS/2” looks like a fraction, this is more childish than clever. While OS/2 had its problems it was a full-fledged operating system which is more than can be said for its early competitor Windows 3.x which required DOS to run and relied on DOS for its file system.
    * The name “OS/2” came late in development, it was originally called “DOS 5” and then “Advanced DOS”.
    * No mention or reference to the excellent book “Inside OS/2” by Gordon Letwin.
    * The PS/2 series turned out to be unsuccessful but they did incorporate some advanced concepts for the time. Also not mentioning the fact that VGA originated with the PS/2 seems like a deliberate attempt to dismiss the PS/2.
    * Windows NT 3.x was definitely not “ridiculously stable”, in fact it was quite buggy. The NT kernel didn’t really become stable until Windows XP. Remember there was the infamous report by Mary Jo Foley regarding Windows 2000 having 63,000 bugs.
    * The primary reason for the failure of OS/2 were lack of apps and perceived cost: OS/2 had few apps and no killers apps like Microsoft did for Windows (Word and Excel) and IBM didn’t promote it to developers like Microsoft did with Windows. As for cost, Windows was perceived to be free since it was included with virtually every PC while OS/2 had to be purchased separately (even on IBM systems). Having DOS and Windows compatibility certainly didn’t cause it to fail. Why would anyone even consider OS/2 just to run DOS and Windows apps in the first place ?

    Having worked at IBM in the mid-1990s, I can say that IBM was definitely much too corporate and bureaucratic, management tended to be stuck with a “mainframe mentality”, they were slow to change direction and they didn’t really understand the consumer market but they did try. IBM made many mistakes but they also had successes, not the least of which being the PC itself. Microsoft wasn’t without its own issues: they gained and maintained their huge market share through preloading which effectively locked out any competition, they never created an API which didn’t have undocumented functions, their software generally took several releases to become stable, their attitude was “embrace, extend and extinguish” and there was a matter of an anti-trust lawsuit against them in the 1990s. Bottom line you can find ample missteps by both companies (e.g. PCjr and DOS 4 for IBM, Windows ME and Microsoft Bob for Microsoft) but an exploration into an aspect of PC history should be more about the product itself and its development and less about old tropes.

    P.S. few people wore suits at IBM in the 1990s outside of upper management and sales.

  7. I chose the title “Half an operating system” quite deliberately. Sure, on the surface it is using the somewhat childish mathematical line that was first used in *.advocacy newsgroups in the 1990s. But that’s just the topmost layer of meaning. An operating system that tries and fails to supplant an existing OS, but is then abandoned, can be said to be half an operating system: it exists, but only in the past. OS/2 was also half an operating system in the sense that half its time was spent running OS/2 native applications and half its time was spent running DOS or Windows apps. OS/2 also had one foot in the 16-bit, 286 world, and the other foot in the 32-bit world: by the time the OS was mostly 32-bit code, it was discontinued.

    In addition, IBM’s execution of sales and marketing for OS/2 could only be charitably described as half-hearted.

    And IBM itself was split in half over the OS. Half of the company wanted it to succeed, and half of it didn’t care, or wanted it to fail. Half of the company was tied to mainframes, and half was trying to take back control of the personal computer market. In the end, the mainframe side won, OS/2 was discontinued, and the entire PC division sold off. IBM today is still half a mainframe company, with the other half being an amalgamation of business software and services.

    The second part of the title is just as important. The triumph AND tragedy of OS/2. The triumph was that it was released and that it was a competitor, and for a brief time it was far better than its competition. The tragedy was that IBM didn’t have the will to keep up the fight, and gave up instead.

    So there you go. That’s the story of the title. I hope it makes more sense now. 🙂

  8. Yuhong Bao says:

    “While OS/2 had its problems it was a full-fledged operating system which is more than can be said for its early competitor Windows 3.x which required DOS to run and relied on DOS for its file system.”
    I mentioned several times how Win9x’s dependency on DOS led Caldera to continue to be able to sue MS.

  9. I don’t think the Caldera lawsuit ended up being that important. The suit was settled in 2000 for $275 million, which was probably less than Bill Gates’ sweater budget at the time. Caldera renamed itself to the SCO Group, tried to sue Linux for a decade and failed utterly, and went from Chapter 11 to Chapter 7 bankruptcy, which basically means it is dead and gone forever.

    At the end of the day, you can’t run a company forever just based on suing other companies.

  10. Yuhong Bao says:

    @Jeremy Reimer: But my point was to add this on top of the (sometimes unethical) attacks MS did on OS/2, which was already bad enough.

  11. dosfan says:

    An operating system is software which manages the resources of a computer and provides services and a application programming interface (API) to run other applications. I’m not sure there is a such thing as “half an operating system” but if anything can be considered a partial operating system then it’s Windows 3.x as it had some operating system features but required DOS in order to run.

    Only OS/2 1.x was purely 16-bit, OS/2 2.0+ dropped 286 support and was 32-bit and only supported 16-bit for compatibility with OS/2 1.x and in VDMs for DOS and Windows.

    I wouldn’t say they didn’t have the will to fight or just gave up, they put quite a lot of time and effort in OS/2. Keep in mind it was on the market from 1987 to 2001. I also wouldn’t say that OS/2 was discontinued because of the mainframe division, it was discontinued because it wasn’t profitable. By 2001 it was clear that Microsoft had won with Windows. Windows 95 and 98 were clear market successes and Windows XP was poised to be an even bigger success (and it was). IBM didn’t understand the consumer market so there were never any killer apps for OS/2, they didn’t promote OS/2 to developers to write apps and Microsoft’s dominance via preloading Windows proved to be a huge hurdle. Add to this the fact that even the IBM PC Company (the hardware group which made their PCs) wouldn’t preload PC DOS or OS/2 on their PCs much to the dismay of the Personal Systems Products group (which developed PC DOS and OS/2), well if IBM wasn’t supporting their own OS then why would anyone else ?

  12. dosfan says:

    @Yuhong, what is the relevance of Caldera to the discussion of OS/2 ? Must you always go off topic ?

  13. Yuhong Bao says:

    @dosfan: My point was to prove MS made a mistake in abandoning OS/2 and later attacking it.

  14. I don’t think you’ve proven that. 🙂

  15. dosfan says:

    Since Windows was successful (until Windows 8.x came along) and OS/2 was not, where is Microsoft’s mistake ? While I don’t agree with many of the things Microsoft has done in the past, you can’t deny their success.

  16. Yuhong Bao says:

    @dosfan: See above and my comments on the Ars article.

  17. dosfan says:

    Again where is Microsoft’s mistake in going with Windows which was 100% theirs instead of OS/2 which wasn’t ? Obviously their decision was successful so it wasn’t a mistake. All your comments illustrate is that you have a bone to pick with Microsoft. No Microsoft wasn’t perfect, yes they did some very questionable things but IBM certainly had its problems as well. Bottom line is that the PC market as we know it wouldn’t exist if it weren’t for both of these companies.

  18. Yuhong Bao says:

    @dosfan: DR was a bigger threat than IBM to MS back in 1990.

  19. dosfan says:

    Digital Research wasn’t a threat to anyone. Their relevance ended with the decline of CP/M once Gary Kidall and friends blew the deal with IBM. The less than 100% compatible DR-DOS is nothing more than a footnote in PC history. Anyway enough of the off-topic discussion, the original posting was about the OS/2 article on Ars Technica.

  20. Dave Yeo says:

    dosfan said:
    -Why would anyone even consider OS/2 just to run DOS and Windows apps in the first place ?

    I started running OS/2 as a DOS replacement. Could run Terminator in one session downloading using YMODEM (no error correction so 1.7 KB/s on a 14.4 modem) and play a game or run a Windows program in another session. Even connected to the internet at first using WinTrumpet and Netscape 1.x with Windows in an OS/2 session. Worked very well, real multitasking with close enough to native speeds and good hardware support.
    After a while I explored OS/2 more and slowly moved to native apps but as I only had 4MBs of memory I never loaded the WPS and missed out on some of the advantages of OS/2.
    This was another failure for OS/2, the price of memory stayed high and IBM’s claims about running in 4MBs was a bit of an exaggeration. It wasn’t until I got 8MBs that I really appreciated the OS/2 desktop.

  21. Paul says:

    dosfan/Jeremy/everyone just ignore Yuhong (wherever he pops up)…its just easier…he doesn’t realise his version of reality is so twisted and unrelated to everyone else’s 😉

    Regarding OS/2…a big part of the “failure” from my point of view was always availability…(growing up in Yorkshire/England) it was incredibly easy to get (legitimately or otherwise *cough*) DOS/Windows and apps/software for them. I always wanted to try and play with OS/2…I came from the Commodore 64 world and something just kinda clicked with me about OS/2. I read anything I could about OS/2. I remember pouring over PCW and other magazines for any info. By the time OS/2 Warp was available as a preview on cover discs let alone anyone I knew using it…it was just too late…Windows 95 (from the beta editions) were everywhere and it was such an improvement over WfWG311 everyone used it.

    My viewpoint aligns quite closely with Michal’s point of view I think, rather than Jeremy’s, but I do agree completely with Jeremy that a lot of this is just opinion so can’t given that caveat either point of view can’t be dismissed. 🙂

  22. Michal Necasek says:

    Microsoft’s behavior is completely understandable… if not ethical or in some instances legal. I think they mishandled OS/2, but they did learn from their mistakes and applied the new knowledge to NT (a huge chunk of the NT team consisted of ex-OS/2 developers). OS/2 at the very least made a difference because Microsoft took it seriously and had to try building something better.

  23. Michal Necasek says:

    Yes, it’s off-topic, but I can’t quite agree. In the late 1980s and early 1990s, Microsoft took DR-DOS very seriously. I’d even say DOS 5.0 was largely a reaction to DR-DOS. Whatever the technical merits of DR-DOS or lack thereof, the fact is that Microsoft was very worried about DR-DOS being widely adopted by OEMs and taking a chunk out of Microsoft’s license to print money. But I highly doubt Microsoft ever considered DR to be a bigger threat than IBM.

  24. I wrote my own post about the history of OS/2 when it passed its 25th birthday. Maybe you’ll appreciate it.

  25. Carlos Osuna says:


    I think that article is faithful to Ars unwritten, but almost official, goal to “promote Microsoft at all cost”.

    Most of it is pure fiction but based on plain truth. IBM acted short sighted with regards to OS/2 goals, requirements and milestones. Both the 386 and the UI should’ve been priorities considering that both Microsoft and IBM had a small window of opportunity before Apple released the Macintosh II, which overcame most of the prior Mac shortcomings. The Amiga, although it had the correct form factor, never got past it’s image of a toy mostly due to the Commodore brand. Same happened with the Atari ST.

    With that in mind, both Apple and IBM went along the same mistaken path of putting legacy before functionality. Microsoft split forces, with Windows/DOS developing parallel to Windows/NT.

    BTW. I don’t think IBM doubted on using the OS/2 moniker since Ars is right that they had a full vision before they had the OS. It was Microsoft that probably wanted to call it PC DOS 5 /MS-DOS 5 to avoid market resistance.

    Today as you mention, the Windows brand has been tarnished by the Windows Mobile, Windows Phone and Windows RT fiascos. Windows has had a hard time transforming into something people view as über-personal. They say “it’s for (the) office and should stay there”.

  26. Brian says:

    The Ars article and this article are both excellent and brought back a flood of memories. Including the angst present with IBM at the time.

    Before running OS/2 at IBM, I ran AIX PS/2 1.1 on my Model 80 (386). Super-fast, ultra-stable, great for software development that targeted the RSIC/6000. After that job was over, I moved to a group doing multi-media on top of OS/2 1.3. After installing OS/2 on that Model 80, it felt as if the 386 had been replaced with a slow 286 and a slower disk installed. But the 1.3 kernel was solid and ran well; the GUI was slow, buggy, and cantaknerous.

    Running CorelDRAW on Windows 3.0 at home vrs. running it on WinOS2 or, more horribly, running Corel’s OS/2 native version really opened my eyes to the bleak future OS/2 faced. Sadly, IBM might have done better: for internal company use only, PMDraw was rock-solid, super fast, and a sheer joy to use. I used to miss it, though the excellent Inkscape has now fully replaced it!

    The comments also reflect some of the angst. IBM’s legendary internal world-wide network hosted many “fora”, what the rest of the world calls newsgroups. Most were relativley calm and civil. But OS2ARENA was a hotbed of name-calling, hair-pulling knock-down drag-out verbal brawls. And I do remember having participating! The CORELDRW forum was chock full of great tips and responive help abot CorelDRAW. However, the CORELOS2 forum was mostly full of unaswered questions and sob stories from folks trying unsuccessfully to get the Windows version of Corel to run on WINOS2. Another eye-opener for me.

    Anyway, thanks to all, authors and comentors alike, for the great memories and good information!

  27. Michal Necasek says:

    When the JDA was signed (1985), it talked about CP DOS. It’s clear that OS/2 was developed as “DOS” of some form (either CP DOS or DOS 5). See all the DosXxx API calls, among others. Hard to say about IBM but Microsoft obviously considered OS/2 to be a future version of DOS during development. How much difference the name made is another question 🙂

    I’d argue that putting legacy before functionality was not mistaken at all, and made Microsoft billions of dollars. However, it’s a strategy that can’t last forever (as we’re witnessing now).

  28. Yuhong Bao says:

    “I think that article is faithful to Ars unwritten, but almost official, goal to “promote Microsoft at all cost”.”
    I have never seen an article on the entire MS OS/2 2.0 fiasco that is what I call complete and detailed and many omitting for example the unethical attacks MS did against OS/2 such as “Microsoft Munchkins” or the original MS OS/2 2.0 SDK from the end of 1989.. I try with my own blog article, but I admit it is not very good either.

  29. Michal Necasek says:

    I always wondered why IBM first almost secretly enhanced AIX PS/2 to run on non PS/2 systems and then canceled it entirely… maybe the RS/6000 folks didn’t like that?

  30. Yuhong Bao says:

    @Michal Necasek: Wasn’t NT originally OS/2 NT?

  31. Brian says:

    Re: Half an operating system. This term was used by many in IBM who preferred Unix or later Linux. One IBM Yorktown researcher’s tag line was”OS/2 is the Integrating Platform. AIX PS/2 is the Differentiating Platform.”

    Re: The will to win. What most folks don’t know is that IBM once hated AT&T with a passion. The PC-RT was a Unix-based minicomputer that was hobbled by IBM to keep Unix from cutting into mainframe markets. As such, most of IBM’s Unix experience was hired from the outside. It was my opinion that the RS/6000 team in Austin couldn’t relocate due to the Texas real estate collapse in 1985. They had no other choice but to make their baby succeed. The RS/6000 was as good as the PC-RT was bad. That it rose to the #2 position behind Sun was a testament to the Austin RS/6000 folks. So I agree with the author’s conclusion that for OS/2, IBM lacked the will. (And I feel fortunate to be one of the rare IBM Boca Raton folks who had the chance to leran Unix on the job.)

  32. “I think that article is faithful to Ars unwritten, but almost official, goal to “promote Microsoft at all cost”.”

    Really? I would have thought that the bit at the end where I compare today’s Microsoft to yesterday’s IBM (calling Microsoft a modern-day legacy company in a world where there is another disruptive transition going on) would have had the opposite effect.

    Surely my Ars overlords would have noticed this and altered the ending?

    But in fact, they altered almost nothing. The article ran 99.9% as I originally wrote it.

    What is hilarious to me is that for every article Ars writes, someone will write a snarky response calling Ars biased in favor of one technology company. Only it changes with every article. Some will post that Ars is totally biased towards Apple. Some will post that Ars is biased towards Google. Still others insist that Ars is pro-Microsoft, and others will scream at the top of their lungs that Ars hates Microsoft and is actively working against it.

    We must be doing something right. 🙂

  33. Pete says:

    I really don’t remember it being called OS/2 for quite a while. It was CP DOS or, more commonly, the internal code names, like Cruiser for the 32-bit version.

  34. Yuhong Bao says:

    @Jeremy Reimer: Well, as I said above, it wasn’t just you.

  35. DOS says:

    Sorry to bring up DR-DOS again, but I for one thought it sounded pretty good compared to the MS- and PC-DOS versions available at the time, only to find out it wouldn’t work with Windows 3.1 and I had to take it back and just stick to MS-DOS. Perhaps without anti-competitive behaviour from Microsoft, DR-DOS could have been popular with those who cared to choose an OS other than what was pre-loaded onto their PCs, just like OS/2, i.e. not really that popular.

  36. Michal Necasek says:

    It was, and except for the name it had nothing to do with OS/2. Different people, different design, different code.

  37. Michal Necasek says:

    Competing in a UNIX market with no dominant player was very very different from competing against Windows in the PC arena. Going against DOS/Windows, it wasn’t enough to have a good or perhaps even better product. In the UNIX world, that alone meant a lot.

    One day I plan to write about IBM’s PC UNIXes… PC/IX, XENIX, and AIX. All developed by 3rd parties and neither made it very far 🙂

  38. Michal Necasek says:

    I didn’t get the impression that the article promoted the 2013 incarnation of Microsoft at all. The 1993 version yes, but that hardly matters today.

    And yeah, if you get complaints of being both pro- and anti-Microsoft, you’re probably doing it just right 🙂

  39. BlahBlah says:

    As a side note:

    pg108-109 of Showstopper! by Pascal G. Zachary tells a story of a MS employee:

    “IBM worshiped hierarchies …
    each employee is worried about his piece of turf and nothing else..
    the (IBM) guy studied (the crash dump screen) and said: `Glad that is not my code` and went back to work..
    At MS this would never happen.”

  40. Michal Necasek says:

    The problem with Showstopper! is that it is strictly one-sided. Of course it says nothing about how unhappy IBM was with Microsoft’s sloppiness and lack of organization. Somehow I suspect that the Microsofties of 1990 would hate Microsoft of 2010 just as much as they hated IBM back then.

  41. Richard Wells says:

    IBM’s problems with bureaucracy yielding code that technically worked according to specifications but did not yield results that purchasers wanted applied to Topview before OS/2 and some projects after. It was an organizational issue. MS suffered the same problems with Vista.

    Unfortunately, with OS/2, having multiple geographically spreadout development sites in multiple companies and no clear single decision maker made it easy to blame someone else for all the problems. Frankly, it was amazing that any useful product resulted.

  42. Yuhong Bao says:

    “The problem with Showstopper! is that it is strictly one-sided. ”
    And it was written before MS did most of the unethical tactics to attack OS/2 later on anyway.

  43. Having read the post of Esther Schindler I replied on her Google+ and wanted to share it here also. Bit long. @Michal, maybe you could do a piece about the boot-manager. I used is for years with full satisfaction! Or did you already write about it and have I missed it? I love your museum and visit it regulary!

    I think it must have been in ’95, I was 17 years old and had finally saved enough money to buy a brand new Pentium. Coming from an Amstrad 286 16 Mhz with 4MB RAM, 40MB HD, VGA and SoundBlaster compatible (which played Dune II quite well, ‘Taking advantage of extended memory…’,I loved those words appearing on my screen), that Pentium was a big leap. I was keen about my choice of hardware, I selected a clone running at 100 Mhz with the motherboard running at the same speed. Paid extra for 128MB of memory. 1GB HD how would I ever fill 1000 MB? I remember the quad-speed cd-rom being noisy and a bit buggy sometimes. For sound I had an ESS card. And of course my first modem, a trust 14k4. Told my mom I was wiring the phone to my room and that I would pay if she thought the phone bill would get to high. Put in a Zipdrive with a simple SCSI card and an S3 virge video later.

    So with the hardware in place, came the choice of software. I wanted OS/2 to come with my new computer. An unusual choice, so I was told. I don’t know exactly but I think European or Dutch court ruled that ‘koppelverkoop’ that is ‘bundled selling’ was illegal and the customer should have a choice. The guy from the store warned me for not being able to deliver my order pre-installed. I remember asking how they would know the system worked, he said they would install Windows95 to check, I don’t know if they ever did.

    So there it was, my custom ordered new computer, plugged in and ready to be booted. The OS/2 Warp with bonus CD-ROM, 3 boot diskettes and a phonebook thick pocket size manual still in plastic on my desk. The most interesting thing I read in this manual about being able to boot from multiple partitions by installing a bootloader. So I could recreate my old installation on a bigger partition of let’s 100MB, having al my software installed and room to share. No more deleting stuff to make this or that that 5 disk game fit. And being sure I could play Dune II.

    So just following the procedure I managed to get dual boot working and I installed Dos by just typing format /s c: on one patition. And OS/2 on the other. It worked! Only 16 colors VGA and no sound. But it worked. It took me quite a while to get it working. And the modem took me weeks for it to work. I even installed Windows 3.11 on my Dos partition to test if it wasn’t broken. It wasn’t.

    So now I had OS/2. And there was the internet. The only packages that where sold were for Windows. And the 5 or 10 gulden (we didn’t have Euro’s back then) software disks at school, all Dos and Windows, everyone said Windows95 was it. And for doing something usefull I relied heavily on my dos-partition. Installing AutoCAD and so on. So finally I decided to do a reinstall of my system with 3 partitions, OS/2, Dos and Windows95. And an extended partition for Data which would be accesable by all 3 systems. I learned Wimdows95 should be installed in the first partition otherwise it wouldn’t boot.

    I managed to go onto the internet with my OS/2 installation. I tried to learn REXX on a blue monday but all in all I had to less practical use for OS/2.

    Still I think it’s very interesting how and why systems with apparently better papers don’t come to market. Why VHS? Why Windows? Yes yes I know software availablity. I’m my own example, in the end I went with the mainstream. But OS/2 once the system of my choice!

  44. Mixed up my memory about the ammount of memory. Pentium came with 32MB and later when I started working I bought a Pentium III 500Mhz with 128MB RAM (and 13GB HD).

  45. Greg Pringle says:

    In March 1994 I was told by an IBM Executive that IBM had come to an agreement with Microsoft to not compete with Microsoft in the general market place. IBM would only push OS/2 to it’s own corporate customers in exchange for a low cost for Windows 95 which IBM would install on IBM PCs. Microsoft had said it was going to raise the price.

  46. Lorne says:

    Just to add to the comments on DR-DOS and Windows 3.x. At the time there was the incorrect thought that DR-DOS was incompatible with Windows 3.x, that you could not run Windows 3.x on top of DR-DOS. This incorrect information came out of Microsoft itself and the AARD code which was put into Windows and was active in a number of beta copies of Windows. The code would check to see if Windows was runnig on top of MS-DOS, or if it was running on some other DOS. If it was found to be running on a competitor’s dos, it would present an error message on the screen implying that Windows would not run properly. There was no real incompatibility, however it lead many to believe that Windows had to run on top of MS-DOS. This has all been well documented in US-DOJ antitrust filings.

    As for OS/2, in the early to mid 90’s it is important to remember that its main competition was viewed to be NT, and when compared with NT at that time OS/2 actually did quite well. 32-bit OS/2 consistently outsold 32-bit NT up until NT 4.0 came out in 1996. Any time the trade publications of the time compared OS/2 with another OS, it was usually NT (and sometimes Novell). In fact in December 1995 IBM announced that OS/2 sold over a million copies in that single month alone, this was apparently more copies sold in a single month then NT sold.

    This was also at a time when IBM was actively trying to market OS/2 in new ways to Win3.x / 95 users. Does anyone remember the Just Add OS/2 Warp campaign? It was aimed at Windows 3.x and 95 users, if I recall IBM won several awards for that campaign. It came with a downloadable app which provided info about OS/2, included a demo, had a “hardware tester” that could tell you if your hardware was compatible, and had a Windows 95 long file name converter and dual boot utility. Of course today a downloadable app is something we don’t think twice about (actually, with all the viruses out there we probably think even more about whether or not to download something like that), but at the time it was very forward-thinking.

    You can still download the files at:

    One thing that is worth mentioning that I think gets overlooked but IMHO owes a lot to the eventual discontinuation of OS/2 (in addition to the items already mentioned in the article and comments) is the failure of the Power PC chip to gain any significant share of the market. IBM spent a lot of time, development effort, and a huge amount of money to get OS/2 running on the PPC. Then as PSP was nearing completion of OS/2 for the PPC (reports suggest that the product was around 90% complete, mostly missing the networking stuff) it was realized that there were very few PPC based computers on the market, and very few companies other then IBM had any plans to release PPC-based computers (aside from the Apple Mac and various *nix based systems). Without the hardware there is no market for the software and the product was cancelled. Of course it could also be argued that because the PPC version of OS/2 was behind schedule, that lead to a lack of hardware. Either way if the time, effort, and development money had been put into the Intel version of OS/2 instead of PPC, I have to wonder if things might be a little different today.

    It is also worth noting that at one time Microsoft had a PPC version of NT (3.5 and 4.0) that has also been discontinued.

  47. Michal Necasek says:

    I’ve written about the PowerPC misadventure before… and I completely agree that the failure of the PowerPC platform was a big blow to OS/2. It’s questionable whether OS/2 for PowerPC (née Workplace OS) would have been a successful product, as it was clearly far behind NT and Solaris in terms of completeness (obvious by comparing where each project stood by circa December 1995). Then again NT and Solaris for PowerPC in the end did only very marginally better than OS/2, since the whole platform collapsed. IBM was definitely burned by that project. The hardware side survived because it was folded back into the RS/6000 workstation line, but the software was useless without a mass market for PowerPCs.

    The whole PowerPC thing was simply a mass delusion predicated on the theory that Intel was a bunch of idiots. But that was not the case. In the end I don’t believe the software could have changed anything.

  48. Craig says:

    Did Microsoft ever attack OS/2? They directed all their attacks against Team OS/2 as far as I can tell. Team OS/2 was not IBM and didn’t own OS/2.

  49. Yuhong Bao says:

    I still consider it an attack on OS/2, and some of these attacks were pretty unethical, and add the DR-DOS stuff mentioned above, and…

Leave a Reply

Your email address will not be published. Required fields are marked *