Tipping Points


One of my birthday presents was the book “The Tipping Point” by Malcolm Gladwell.

In it, he talks about how epidemics and other flash occurances happen – situations that are stable, and a small thing changes and suddenly the world changed overnight.

I’ve been thinking a lot about yesterdays blog post, and I realized that not only is it a story about one of the coolest developers I’ve ever met, it also describes a tipping point for the entire computer industry.

Sometimes, it’s fun to play the “what if” game, so…

What if David Weise hadn’t gotten Windows applications running in protected mode?  Now, keep in mind, this is just my rampant speculation, not what would have happened.  Think of it kinda like the Marvel Comics “What if…” series (What would have happened if Spiderman had rescued Gwen Stacy, etc [note: the deep link may not work, you may have to navigate directly]).

“What If David Weise hadn’t gotten Windows applications running in protected mode…”[1]

Well, if Windows 3.0 hadn’t had windows apps running in protected mode, then it likely would have not been successful.  That means that instead of revitalizing interest in Microsoft in the MS-DOS series of operating systems, Microsoft would have continued working on OS/2.  Even though working under the JDA was painful for both Microsoft and IBM, it was the best game in town.

By 1993, Microsoft and IBM would have debuted OS/2 2.0, which would have had supported 32bit applications, and had MVDM support built-in.

Somewhere over the next couple of years, the Windows NT kernel would have come out as the bigger, more secure brother of OS/2, it would have kept the workplace shell that IBM wrote (instead of the Windows 3.1 Task Manager).

Windows 95 would have never existed, since the MS-DOS line would have withered and died off.  Instead, OS/2 would be the 32bit application for lower end machines.  And instead of Microsoft driving the UI story for the platform, IBM would have owned it.

By 2001, most PC class machines would have OS/2 running on them (probably OS/2 2.5) with multimedia support.  NT OS/2 would also be available for business and office class machines.  With IBMs guidance, instead of the PCI bus becoming dominant, the MCA was the dominant bus form factor.  The nickname for the PC architecture wasn’t “Wintel”, instead it was “Intos” (OS2tel was just too awkwards to say).  IBM, Microsoft and Intel all worked to drive the hardware platform, and, since IBM was the biggest vendor of PC class hardware, they had a lot to say in the decisions.

And interestingly enough, when IBM came to the realization that they could make more money selling consulting services than selling hardware, instead of moving to Linux, they stuck with OS/2 – they had a significant ownership stake in the platform, and they’d be pushing it as hard as they can.

From Microsoft’s perspective, the big change would be that instead of Microsoft driving the industry, IBM (as Microsoft’s largest OEM, and development partner in OS/2) would be the driving force (at least as far as consumers were concerned).  UI decisions would be made by IBM’s engineers, not Microsoft’s.

In my mind, the biggest effect of such a change would be on Linux.  Deprived of the sponsorship of a major enterprise vendor (the other enterprise players followed IBMs lead and went with OS/2), Linux remained as primarily an ‘interesting’ alternative to Solaris, AIX, and the other *nix based operating systems sold by hardware vendors.  Instead, AIX and Solaris became the major players in the *nix OS space, and flourished as an alternative. 

 

Anyway, it’s all just silly speculation, about what might have happened if the industry hadn’t tipped, so take it all with a healthy pinch of salt.

[1] I’m assuming that all other aspects of the industry remain the same: The internet tidal wave hit in the mid 90s, computers remained as fast as they had always, etc. – this may not be a valid set of assumptions, but it’s my fantasy.  I’m also not touching on what affects the DoJ would have had on the situation.

Comments (84)

  1. Chris says:

    In a Windows vacuum, OS/2 would not have been as successful as Windows, for various reasons, and it’s possible that the PC architecture wouldn’t have quite the overwhelming dominance that it does.

    What happens to marginal players like Apple, Digital Research, NeXT, Be, each of whom could have been a contender on the PC — OR, in a less-PC-dominated world, more successful with their own hardware efforts.

    It’d make a great alternate history novel.

  2. Chris,

    That’s totally possible. It’s not clear if the delay of 4 years until multiple DOS applications became available on OS/2 (and the revitalization of MS-DOS that Windows enabled) would have allowed alternative OS’s to be far more successful than they were.

    It would make a great novel 🙂

  3. bramster says:

    From the dark side. . .

    Would a "Microsoft Flight Simulator" have been availbable to Mohammed Atta?

    Each "what if" root has many branches

  4. Needless to say, as a Linux user I dispute the claim that Linux wouldn’t have succeeded without IBM.

    I’ve been using Linux since well before it was conceivable that IBM might even *use* it, let alone back it to the tune of a billion dollars. And it was clear well before IBM’s involvement that Linux’s success was not just a clear possibility, but inevitable. If IBM hadn’t been the first Big (blue) Fish to realize this, someone else would have.

    Furthermore, it seems likely to me that if IBM had been the driving force behind the primary PC operating system, they’d have made a much bigger point of trying to shut clone vendors out, probably with great success. The PC’s success as an architecture is *entirely* due to widespread hardware competition. If that hadn’t happened with the PC architecture, I suspect that some other cloneable architecture (with a different OS) would have filled the interoperability vacuum, but if *all* the vendors had stuck with "don’t clone me" attitudes, the difference from today would be far more radical. It’s hard to imagine what the world would have been like with a wide variety of different hardware and operating systems *all* in broad use, but it seems to me that in such a world, a new OS like Linux would have a *better* chance of getting a foothold (since it’s entering a market that already has competition, rather than trying to overcome a monopoly).

    Application vendors, for example, would probably have put a much greater premium on portability from day one, meaning porting to the new up-and-coming OS would be a no-brainer instead of a massive amount of investment.

    I suspect Sun might have ended up with the monopoly when they came out with Java (making portability automatic), unless someone else had beaten them to it.

    It’s all very interesting speculation, anyway 🙂

  5. Stuart, you might be right – I’m figuring that the "Linux revolustion" has been fueled by two things:

    #1: Dominance of the PC platform – that doesn’t change in my scenario.

    #2: Corporate interest in Linux as a platform. And that does.

    Because none of the major consulting firms would be pushing Linux, instead they’d follow IBM’s lead. The other hardware vendors (Compaq, HP, Sun, etc) would continue to sell their own brand of *nix – they’d have no interest in Linux as a platform, so they wouldn’t be pushing it.

    So the only people pushing Linux would be the FSF types – there’d be little corporate support except as a replacement for other *nix platforms.

    On the other hand, many of your other points (especially regarding app portability) do make a lot of sense.

  6. I guess I put it the other way around: the corporate interest in Linux was fueled *by* its undeniable technical and grassroots-level adoption success.

    Remember that in the real world IBM picked up Linux despite having its own Unix brand. Linux beat out IBM’s best efforts (AIX and the stillborn Project Monterey) on *merit*, so convincingly that IBM themselves decided to scrap their own work in favor of it. I have a hard time thinking of any corporate involvement (on the scale you’re contemplating) before that point that could be said to explain IBM’s decision to adopt it. So I’m forced to conclude that if not IBM, one of the other hardware/Unix vendors would have done what they did. The other hardware/Unix vendors, in the no-Windows scenario, would be in the same place that IBM was in today’s world, with the same options available.

    I’d definitely add one to your list of things that fueled Linux’s success, although it doesn’t affect the "what if" because neither of our future-histories modify it: the widespread availability of the Internet. Linux is an (IMHO inevitable) product of the fact that suddenly anyone with programming talent can easily get the latest version, submit a code patch, and see it integrated into new versions within days, if not *hours*. Linux couldn’t have happened if the developers had to mail around 3.5" floppies 🙂 My guess is that the absence of the Internet is pretty much the only thing that really *would* have erased Linux out of history.

  7. greenlight says:

    Another interesting point is if IBM had a bigger stake in the PC business, would they be as interested in joining the AIM (Apple/IBM/Motorola) cooperation, and working on PowerPC?

  8. SuperBK says:

    The dominance OS/2 would have changed my life quite a bit. We invested serveral years in OS/2, starting in 1989 with 1.3. This was the next step for our "system" after DOS. We continued to develop for it until 1995, or about until the beta of NT 4.0 came out. Then while still shipping OS/2 systems, we spent the next 2 years porting everyting to NT. The experience paid off in the sytem areas: threading, inter-process communications, TCP/IP socket programming, etc. We had to rewrite the GUI wich was mostly text based anyway. It seems to me that that with Microsoft’s resources behind it, we would still be using OS/2 today. IBM should have released it to the open source community, but that may not have been possible because it contained some code owned by Microsoft.

    Things might have been so different I might not even live where I do now. Talk about life changing events!

  9. Lee says:

    It’s easy (and fun) to play "what if," since every situation you use it on had some critical path of events that directly enabled it. Bramster alluded to this earlier, as have the pens of many great authors.

    What if David wasn’t able to get Windows applications to run in protected mode? Well, perhaps somebody else may have. Perhaps he would have found some other, better, means of memory magic. Perhaps David’s discovery actually prevented somebody else from finding some other, better, means of memory magic.

    For any positive experience in life, I can usually look back and find an absolutely terrible experience that was fundamentally crucial in enabling that positive experience.

    Take your significant other, for example. All of those amazing experiences were directly enabled by some tragic breakup with an ex-girlfriend.

  10. Cheong says:

    I think the effect might not be that great.

    It’s true that the Protected-mode development may have to be delayed. But since I believe David is not the only person who know about protected mode things, Microsoft may be able to find other, while possibly won’t be able to write those efficient codes.

    Microsoft once have been inspired by the GUI environment, I don’t think it’ll so easy to give it up. They’ll continue to hire people until they find someone who is capable to handle the project.

    And Windows may behave differently. Try taking the weight of that significant person out and add the ideas of another person in will quite probably cause a series of branches in the timeline, and the output should be quite different.

    For me it could be interesting to think about "what if Microsoft hire Linus to work for them at the time of initial release of Linux kernel?" 🙂

  11. Cheong,

    You’re missing the skunkworks aspect of the project – Nobody in the systems division (except for David and the rest of the Win3 team) wanted to get windows applications working in protected mode – it would directly compete with Windows and Microsoft’s strategic direction was 100% focused on OS/2. Windows was a distraction.

    If David hadn’t done it, then nobody would have – that’s why I described it as a tipping point. If David hadn’t been working on that project at that time, Windows 3 wouldn’t have happened and history would have changed.

    Oh, and Lee, I didn’t have a girlfriend when I met Valorie 🙂 Now there was a tragic breakup with HER boyfriend, but not mine 🙂

  12. David Candy says:

    What if … IBM didn’t approach MS.

    Microsoft, whose first OS they sold was Unix, would be babbling about Unix’s stability, scalability. (I just love pointing out to anti MS that MS was unix company).

    But where are the microprocessors. My company bought Dos 3.3 and DBase III+ (and an AT to go with it) because it would have cost millions and taken years with thousands of plane trips to coordinate it to allow my division (11 people) out of 60,000 people to do mailing labels in lower case (and like that was going to happen). Then our accountant discovered 3270 terminal hardware for the PC and so he got one for Lotus 123.

    NT is more mainframe like (in terms of the overall business system). But what do little divisions of 11 people (who competed with our 60,000 strong parent – they didn’t want us but had to buy us as well if they wanted a particular TV factory) if they need IT outside the network structure.

    MS gave us empowerment with DOS, then took it away with NT.

    If I was MS. I’d get WinCE on turnkey systems (but desktop sized). Sinclair tried this and it failed. My dad can use his digital camera (an operating system – and I can’t really use mine) but not XP. Security problems – what security problem – you can’t d/l programs. For a lot of people you need function based computers – this thing prints digital pictures, can make letters, can web browse, and can send recieve mail, and that’s all it does. My camera asks me no questions and claims it can print (I’m all electronic). This is what is needed. Who asks questions about the OS in a TV, whose TV has ever crashed? Yet we get to see NT stop errors on our public transport scteens instead of when the next train is. OS/2 ATMs also crash, especially after sabotage.

    While guys n gals like me wouldn’t buy it in a pink fit. My dad would, especially if it doesn’t have the word computer in it. My mum might buy it (she is computer literate). She prefers turnkey systems.

    Anyway OS/2 is alive and well and updated. It seems to be ideally suited to playing multimedia ads on ATMs (our company did usability testing on one such system – we had ATMs everywhere and in every office. Then they started getting sabotaged as people who had to share an office with a multimedia ATM playing ads every 2 minutes couldn’t handle it anymore)

  13. David Candy says:

    Oh, and Lee, I didn’t have a girlfriend when I met Valorie 🙂 Now there was a tragic breakup with HER boyfriend, but not mine 🙂

    I’m tempted. You can’t do things like this and expect me to resist.

    I hope you coding does what you mean it too. Your writing doesn’t say what you mean it too.

  14. jason says:

    os/2 is a resource hog.. it does what it does with resources efficiently but it needs them all.. if David Weise didn’t figure out what windows needed someone else would have

  15. NSK says:

    Very interesting post, but I don’t think that the lack of Linux support from IBM would be enough to cause it to stay just an alternative to the other *nixes. Linux would eventually compete with OS/2 and perhaps replace it more easily than Windows.

  16. Wired reader says:

    Cheong wrote: For me it could be interesting to think about "what if Microsoft hire Linus to work for them at the time of initial release of Linux kernel?" 🙂

    There is an interesting piece on what happens if they do so now (as a memo by Linus to BillG, in retrospect from 2008 ;-)) at the above URL.

    One of its true gems is this:

    I was stunned when you agreed to accept the General Public License mandating that everything you added at the level of the new operating system would remain open. But you’ve been true to your side of the bargain, and you’ve won my respect. You never made me alter my goal, which was world domination for Linux. I’ll never forget your line: "Come on, Linus, infect the mothership." I still believe that was the best recruiting pitch ever uttered. We both took a lot of criticism from our partisans, but look what we’ve accomplished. The world is using software that doesn’t suck!

  17. Mike Tran says:

    Don’t want to be a tech nerd but we would most prob be running OS/2 5.X as the last os/2 out was os/2 Warp 4.0 Connect 🙂 (old school os/2 fan here)

  18. Bif Naked says:

    I don’t think that IBM would have pushed OS/2 into the server realm. They already other server OS’s and there would be no point in competing with themselves.

    I think if MS had failed and OS/2 had won, then it would have been OS/2 on the desktop and unix on the server, but I wonder if we would have gotten to the inevitable unix everywhere any sooner. Would the FOSS community have been as keen to create Gnu/Linux if they didn’t have to overcome the monumentally bad server os, NT? The early adoption of Gnu/Linux was driven by people who simply couldn’t get NT to run a web server for any length of time without crapping out. They were willing to overcome any learning hurdle.

    The whole windows path has been a detour in the eventual progression to unix everywhere. Not because unix is a better OS, but because unix isn’t a single entity. Unix is a way of creating an operating system that is based on standards and produced in a cooperative fashion. The superior way in which it is developed is why everything will eventually be unix based.

  19. Seegras says:

    Having OS/2 backed by IBM and dominating the PC-market and Microsoft doing things like Microsoft Office/2 could have lead to other interesting things:

    – NeXTSTEP still alive. As something like an "upper-class" Apple.

    – Compaq lacking the force (because of stricter control of the PC-architecture by IBM) to overtake HP and DEC; making those two still alive; and making Compaq a PC-manufacturer and nothing else.

    – Linux steamrolling mainly the unix-territory, breaking into SGIs, HPs, DECs, Suns (and partly IBMs) marketshare and eventually getting backed by one or several of those. Not likely by IBM, though.

    – Novell still going strong with Netware.

    – Maybe even with Commodore as a competitor in the home-market?

    I think a more diverse marketplace would have been the result in the first place.

  20. Zach says:

    One thing this fails to acknowledge is how MS does business. Even without the tipping point it would only have been a matter of time till MS wanted their own OS to push. And MS is VERY good at getting into markets and taking them over. I have no doubts they would have suceeded in owning the Desktop OS market either way.

  21. mschaef says:

    This is pretty interesting to speculate on.

    If you take the "schism" to be the driving force behind the Apple/IBM alliance, that leaves Apple in the early 90’s with no IBM involvement in ‘Pink’ and no next generation CPU. I don’t know where that leaves Apple, but likely, they continue with Motorola 68040/68060 machines, and maybe get forced into alliance or merger with somebody like Sun. SparcMac, anybody? 🙂 (An Apple/Sun alliance would likely have kept Sun from trying to partner with NeXT on OpenStep, which might have hurt NeXT’s would-be credibility badly enough in the high-end (finance, etc.) market to send them under…)

    On the other hand, the reason behind the IBM/Microsoft schism, Microsoft’s ambition, hasn’t exactly gone away in this alternate timeline. My hunch is that the alliance continues with Microsoft fighting all the way for better terms, more ownership, and more automony. Given their history in the PC industry it seems safe to assume that IBM drops the ball somewhere along the way and Microsoft runs with it. Maybe this takes the form of some kind of fork in OS/2 in the late 90’s. The big difference there would be we’d all be coding to some variant of the OS/2 API rather than Win32. The other possibility is that when the internet comes around Microsoft, owning the whole browser, but not the OS, does to IBM what they feared Netscape was trying to do to Windows: undermine the value of the OS.

    No matter how you slice it, life would have been more difficult for MS.

  22. Someone submitted this to slashdot. Some comments there might be interesting…

  23. OS Observer says:

    Windows/386 2.x was available long before Windows 3.0 (and its 286 protected mode). It’s hard to say whether Windows 3.0’s 286 protected mode was all that important in the end. A lot of people thought so at the time, but with hindsight I don’t think it would have mattered much.

    Remember that Windows 2.x also supported EMS and UMBs, and it could use XMS for things like disk cache. So even on a 286 it wasn’t in terrible shape. (Well, in 286 terms. :-)) The "286 PC with lots of non-EMS RAM" was a very small niche (as IBM discovered, despite what its big customers, who insisted on 286-friendly and thus DOS-hostile OS/2 1.x, said). If you had more than a couple megabytes, you probably had a 386. There were also vendors who’d sell you a little chip kit to drop a 386 into a 286 socket. I think even Intel got in on that act.

    Windows 286 protected mode certainly wasn’t reliable, so that wasn’t an advantage. (OS/2 1.x was, particularly in the 1.3 era when it was IBM’s baby, unless you turned on and used the DOS Compatibility Box a.k.a. the DOS Penalty Box. Then you’d get a very weakly compatible, single DOS session which could take down the whole system. So you might as well run crappy Windows.)

    My guesstimate is that 3.0’s 286 protected mode probably hastened the eventual outcome, but that’s about it.

  24. IBM was the 300lb gorrilla. The reason many poeple prefered Microsoft products was because of their anti-IMB attitudes. Yes OS/2 would have replaced DOS but Microsoft, actually the Digital team, would still have produced NT. People would have embrassed it as the alternative OS. NT Windows would have displaced Linux and BE/OS in the Anything but IMB alternative OS timeline. Game makers, in my oppinion the real drive behind PC sales, would be producing DirectX and RAW (Rexx Accelerated Windows) compatible games so we NT users could play a few games on our alternative platform. Microsoft would have joined with Netscape to produce a (Common Network Interface).

    Oh look! The glue on the bottom of my bag is drying. More later!

  25. Alllen says:

    As a user of OS/2 with the OS/2 screen reader written by IBM, I can say that the accessibility for the blind in OS/2 was far superior to Windows, even to this day. If I had only known what David was up to I’d have tried to stp it somehow–now where is that terminator and time machine when I need it?

  26. Alllen says:

    As a user of OS/2 with the OS/2 screen reader written by IBM, I can say that the accessibility for the blind in OS/2 was far superior to Windows, even to this day. If I had only known what David was up to I’d have tried to stp it somehow–now where is that terminator and time machine when I need it?

  27. KenH says:

    Can we even assume the whole GUI thing would have happened without Windows 3.0? At least at that time?

    One of the things that made it attractive was that it didn’t replace DOS, it ran on TOP of DOS. As essentially a graphical shell, it was substantially simpler and cheaper than a full-blown OS would have been, and gave users a bridge between the text and GUI worlds. A quote I heard frequently back then was "well, for $40, if nothing else, it’s good for playing solitaire." How many copies of Windows 3.0 did solitaire sell, I wonder. The only real "native" apps in the beginning were WYSIWYG ones like Pagemaker and Ventura Publisher, and even those were ported over from the Mac. One of them even took a detour through DR’s GEM GUI first.

    I think it’s possible OS/2 might have been positioned where it was too expensive for the home user, and text-mode multitasking products might have reigned at least for a while.

    I still recall one columnist (Still writing for PC magazine, IIRC) predicting that IBM’s "Topview" would be the platform of the future.

  28. IN SOVIET RUSSIA, WINDOWS APPS RUN YOU!

  29. Dutch Rapley says:

    it’s fun to speculate. It could have essentially changed the processor industry too. Where would Intel be without Windows. I’m sure IBM would have pushed RISC processors instead. May companies can give credit to MS for their success.

  30. David Relson says:

    OS/2 was far more capable than Win3.1.

    As an OS/2 user I was accustomed to conveniences like simultaneously typing and diskette I/O. I was shocked to discover that Windows couldn’t multitask like that — with the floppy drive in use, editor keystrokes were delayed. When I saw Windows 95 my thought was "At last — multitasking."

    The workplace shell was far in advance of the Windows desktop. Years later Windows still had the same clunky old Control Panel. Not good.

    Then there are the programming APIs. Prefixes like Dos, Kbd, Dev, made it apparent which OS/2 subsystem included the API function. Windows had (has) thousands of API functions, but there’s a lack of organization, i.e. naming functions.

    ‘Tis too bad that IBM’s marketing was so awful. Their product was supperior.

  31. Wilson says:

    Humm,

    I think there could be even more radical changes…

    IF Microsoft give up the Windows/DOS Line after windows 3.0 but before windows 3.1… they give up the dominant OS(sic) for the next years… and right Windows 95 disaperars…

    BUT I don’t see OS/2 taking the crown in full… OS/2 would be the major OS for x86 PCs but… the others would have a run too… Macs, Amigas, Alternate DOSes, Alternate GUIs for DOS and Alternate OS like Linux…

    I think the PC world would be more mixed…

    And that would mean the end of the wintel lock-in. The CPU race could have started earlier and had stronger results… 64bits Aplha Digital CPUS here available in 1992… if there was no wintel lock in by then… why would the world wait more 10 years just for a x86 compatible64 bits CPUs to start to migrate to 64bits? GPUs in top nvidia and ATI cards have now 256bits… ours main CPUs could be now just like that

  32. Jay Denebeim says:

    Personally, I think it would have gone a bit different. I don’t believe IBM was ever interested in the home market. They would not have pursued it as agressively as microsoft has.

    My vote for which personal computer would have come out on top is the Amiga. It had a fanatic following and if the environment was a bit different I believe it would have done very well in the home arena. About that fanatic following, I jumped to Linux as did most of the other Amiga types (probably the most famous is Matt Dillon, although there are others) at the same time. The reason for this was Commodore tanked at the same time that Linux became usable.

    This added a large active developer base to Linux. People who already knew how to program on a multi-tasking operating system.

    So, if the Amiga hadn’t died I’m not convinced Linux would have taken off.

  33. vince says:

    It always amusing me how MS people can’t really grasp the real concept behind Linux.

    Linux has ben around for as long as NT, and was created because of the lack of a cheap UNIX derivative for commodity hardware. Having a super-OS/2 around would not have changed this at all.

    A better speculation, if you wanted to eliminate Linux, would be to wonder if 386BSD wasn’t tied up with the AT&T lawsuit if Linus would have bothered to start Linux. In that case possibly FreeBSD would have taken over Linux’s eventual role.

  34. CowboyNeal says:

    You’re ignoring the most important question of all!

    WAS WINDOWS 3.0 DIGITALLY SIGNED?

  35. dave levitt says:

    I’m a little confused. I recall ‘protected mode’ operation on i286 machines with various DOS extenders such as DesqView and protected mode operating systems [NetWare/286, Xenix, OS/2 1.x] dating back to 1985 or so [and bank switching memory / overlay extenders back to CP/M days].

    DesqView/386 was capable of executing multiple, protected mode DOS programs [Including a session of Windows/286 and its programs]. A principal Windows 3.0 ‘enhancement’ was breaking this compatibility.

    Developer releases of Microsoft OS/2 2.0 also had the capability of running multiple DOS programs – a principal goal was removing the OS/2 1.x ‘penalty box’ for running legacy software. This support was not very tunable – it was little better than what is available in WinXP today.

    IBM’s flavor of OS/2 provided not simply virtual DOS sessions, but virtual machines capable of running leagcy [like CP/M-86] or protected mode software [like NetWare/386].

  36. Lee Reynolds says:

    I love all the speculation that tries to tie the success of the PC to Windows, or tries to say that somehow someway other platforms would have become dominant without Windows 3.0

    The truth is that the PC was already a screaming success in 1990, and I’m not talking about the "IBM PC," but the PC as a generic platform based off of commodity components from multiple vendors. The "clone" industry was very well established in 1990. So well established in fact that every month a magazine as thick as a college textbook was put out with almost nothing but ads for PC hardware. This magazine was called the Computer Shopper, and it was how one found deals on mail order computer equipment before the days of froogle and pricewatch.

    If Windows 3.0 was addled by the 640k barrier, then it still would have shipped and it still would have sold. Microsoft has a long history of making winners out of products that are technical losers. OS/2 at that point was a DOG. IBM was still trying to play the "It’s our ball and you can’t play with it" game when it came to the PC, and OS/2 was one of their core strategies. The other being the PS/2 line with Microchannel. Both were non-starters. OS/2 because it sucked, and the PS/2 because it was overpriced and incompatible with standard PC’s. I’ve heard before that corporate america bought a lot of PS/2’s, but that is hardly enough to create market dominance.

    By the time that IBM fixed OS/2 with the 2.0 release, Windows was already dominant, thanks in no small part to OS/2 1.x being broken and IBM charging big bucks for it. By the time IBM woke up, the game was already so far along that they only had a small chance of eventually winning. OS/2 3.x was their last best hope for victory, and it fell to Windows 95. Or I should say that IBM gave up when windows 95 was released. I used to use OS/2 3 back before I discovered Linux, and it was a hell of a good OS for its time. I could multi-task DOS applications and even use virtual memory under dos to compensate for only having 6 megs of ram. When Windows 95 came out, I was not impressed with it. They eventually got it right with NT 4.0, but it was a long time coming.

    In any case, PC’s would still be dominant no matter what happened with Windows 3.0, and the only way that OS/2 would be a winner is if IBM woke up and started being a player instead of trying to be a dictator. Without pressure from Windows, it is likely this would not have happened as quickly as it did.

    OS/2 might still have "won," but only due to lack of competition. One can only hope that the improvements that were made to the 2.0 release and up would have still been made. If OS/2 didn’t win, then we’re back to DOS again, at least until someone came up with a viable replacement.

  37. Kent Tegels says:

    Funny that nobody has mentioned that Xenix could have become the defacto OS if both OS/2 and WinNT flopped, giving us a very different world indeed.

  38. JeffC says:

    You are all forgetting about the Amiga. The PC’s real challenge in the late 80’s to early 90’s was from Commodore. They had significant saturation in the home at the time and were making a play for business desktops. If Windows didn’t become a hit, I think we would all be staring at a red and white checker ball at startup instead of an IBM logo.

  39. "Oh, and Lee, I didn’t have a girlfriend when I met Valorie 🙂 Now there was a tragic breakup with HER boyfriend, but not mine :)"

    So you have a boyfriend too? =)

  40. Shayne says:

    Just as a follow up, Microsoft didn’t believe that Windows *could* be PM at that time.

    David did it, literally, as a bet, as I recall.

    When it was working, they went to a management meeting, and said something like ‘We should reexamine the idea of Windows Protected Mode. Oh, and I should say that we’ve got it working downstairs.’

  41. OS/2 2.0 actually came out in 1992, nearly a year before NT. In your scenario it came out in 1993. Are you saying that having MS involved would have actually SLOWED DOWN the project? I believe it! But there’s no way OS/2 would have been only 2.5 in 2001! Even IBM alone added enough bells and whistles to come up with 3.0 in 1994 (integrated TCPIP/internet stack and apps, smaller memory footprint, PnP for PCMCIA, improved GUI) and 4.0 in 1996 (full PnP, easier install, better networking, GUI updates, Java integration).

  42. Jim Drash says:

    Intersting thoughts. I goes to show that much of what we take for granted we caused by "incidental" events.

    What if IBM had not published the BIOS assembly listing for the PC? – No Clone business for years or perhaps decades

    What if Don Estridge had not died in th plane crash in Dallas? – Who knows, we lost a might visionary we he was killed.

    Clearly, there are some major watershed events in comptuing within that last "few" years.

    1) The Emergence and dominance of the PC

    2) TCP/IP as "THE" standard

    3) The Web and the browser

    There are a few more but these are pretty big.

  43. Loser says:

    This is all pretty dumb. MS-DOS wouldn’t have "withered".

  44. Loser,

    Actually, from the OS group at Microsoft, MS-DOS was on life support…

    Stephen, Hmm.. I don’t know, I was guessing on the dates.

    Shayne, Actually I asked Dave about it on Tuesday – the original post has his version of the story (he let Steve know ahead of time and told his managers before the meeting).

    Lee: Yes, the PC as a platform was a screaming success. And MS-DOS was the successful operating system for that platform. And Microsoft, IBM and all the other OEMs were trying their best to get every user switched to OS/2..

  45. Ted Warner says:

    quote: "Intos" (OS2tel was just too awkwards to say)

    what about "ostel" (possibly pronounced AHZtel)?

  46. Ted, I actually thought about using "ostel" when writing this up, but decided that people would say that I was being self agrandizing (since ostel is suspiciously close to osterman :~) ).

  47. Chris says:

    Lee, the other thing is that Windows was obviously not the only graphical shell available for the PC at the time. Digital Research might’ve had more luck with GEM, or at least pushed DR-DOS forward, since Windows incompatibilities were apparently what killed it. Later on, NextStep could’ve been more successful on the PC if there wasn’t Windows dominance; it’s clear that BeOS would’ve actually been available on some name-brand vendor PC boxes as an option if not for MS dominance.

    Also, Apple was keenly interested in making the Mac toolbox portable for their own new hardware efforts. Apple’s StarTrek project (x86) might have been greenlighted — or it might not have happened in the first place if not for the success of Windows 3.0 — which would have had some impact on Apple’s future OS plans, because the high-level portions of PowerPC system software and Copland Mac OS efforts came directly from StarTrek.

  48. RichB says:

    Shayne/Larry, The barbarians book I mentioned in the comments to the previous blog entry goes into more detail with regards the Weise/Ballmer conversation and also the meeting later the same day.

    Perhaps the best quote is:

    "What do we tell IBM?" Ballmer asked.

    Gates hunched over and leaned toward Ballmer. "I don’t know, Steve. That’s your problem"

  49. Someone on Slashdot thought my comment was good. I don’t know whether to feel complimented or insulted!

    (I wanted to reply on slashdot but I can’t connect to it right now – maybe the link from here slashdotted slashdot? 😉 )

  50. mschaef says:

    "OS/2 2.0 actually came out in 1992, nearly a year before NT. In your scenario it came out in 1993. Are you saying that having MS involved would have actually SLOWED DOWN the project?"

    _MS_ OS/2 2.0 was being demoed to journalists before the schism, multiple DOS boxes at all. I remember pretty clearly articles in Byte magazine describing demos given. The "public" story back then was that OS/2 3.0 (aka OS/2 NT) was the next generation portable OS/2.

    My understanding was that at the time of the schism, IBM wanted to differentiate OS/2 2.0, so they spent a lot of time and energy on things like SOM and the Workplace Shell. Both of which delayed the release. OS/2 1.3, was the first IBM release and an interim effort to attempt to solve some performance issues, etc.

    Meanwhile, Microsoft retasks the OS/2 NT project to develop WinNT, the primary difference being Win32 and the corresponding subsystem.

    Look at it that way, and we might all be running variants of the OS/2 NT codebase, which, in a way, we actually are anyway. The IBM/Unix angle is still something of an unknown.

    PS: IF I got anything wrong with the pre-history of OS/2 NY and Win NT, I think our host is likely more than able to correct me (which I hope he would ;-).

  51. mschaef says:

    "Just as a follow up, Microsoft didn’t believe that Windows *could* be PM at that time. "

    I thought I had heard that part of the reason OS/2’s "GDI" layer differeed from Windows’ was that IBM wanted graphics calls that more closely matched their maninframe systems. I always thought it was a shame that there wasn’t more compatibilty between the two systems.

  52. Axinar says:

    An even bigger "What If?" of course is — what if Linux Torvalds had been aware of the existence of the BSD’s in 1991???

  53. Eli Lilly says:

    =====

    What if David Weise hadn’t gotten Windows applications running in protected mode?

    =====

    Then perhaps someone else would have?

  54. brian says:

    I think that Larry’s probably right. With freshly-chastened IBM in control, the de-facto standard OS/2 would have been of better quality, and would not have pursued the illegal anti-competitive strategies that Microsoft did. Whether this would result(as many of us believe it would) in better quality, variety, and economy in software in general is debatable, of course, but it may have been more difficult for Microsoft to use its OS dominance as a weapon against application developers.

    The free-software movement would thus probably have remained among the elite programmers who were mostly upset with the lack of freedom of Unix, which was not so relevant to the masses. Not until it became painfully obvious that the dominant platform in computing was controlled by a megalomaniac with a strong desire to maintain and expand his control to the detriment of all did anyone regard "free software" as anything but an esoteric philosophical notion as obscure and irrelevant to the average person as Unix was then, and would undoubtedly remain in that scenario.

    Of course, one thing that would not have changed is the primitive economic principle that in a competitive market, prices always move toward unit costs. Since the unit cost of software approaches meaningfully close to zero, something would have to give. In a competitive world, it may have become plain far sooner that software as a salable product was not a sound business model (as indeed most people believed before Microsoft).

    If Larry’s speculation is right, however, it really underscores a perversely ironic thought that has occurred to me lately: that far above the Stallmans, the Torvaldses, the Ritchies, Thompsons, and Raymonds, I think that the Lifetime Achievement Award for the Advancement of Free and Open Source Software rightly belongs to William H. Gates III of Redmond, WA. And I think that the time is approaching when – in a gesture far more morally pure and deeply satisfying than that of the lucky fellow who managed a pie in his face – the Open Source community should publicly present this award to the one who so richly deserves it.

  55. Eli, read the other post – there wasn’t anyone else to do it – it was a skunkworks project, and it was only because of David’s unbelievable talents that it happened.

    Brian,

    An interesting viewpoint. I don’t happen to agree with you (not surprisingly), but an interesting point nonetheless.

  56. As much as I loved OS/2, hate Microsoft and love Linux, there would have been some massive problems with IBM at the steering wheel.

    #1: IBM was a failure with the "home market". They had no idea how to talk to anyone except corporate decision makers. Remember the OS/2 Warp ad campaign? You’d have thought it was promoted by IBM’s competitors. It was seriously destructive! Microsoft created a really "user friendly" facade that people could relate to. (Nothing like Apple’s, but that’s another story.)

    #2: OS/2 still had some fundamental problems, especially with the compiler. C/Set (later Visual Age) had some enormous compiler bugs. I spoke with a friend who told of tales that convince me that IBM’s people were completely unable to handle the needs of the growing market.

    #3: There were some "idealistic" programming concepts (CommonPoint, Taligent, OpenDoc) that sounded nice on paper, but developers today understand many fundamental design flaws. I’m not claiming that Microsoft is anything beyond lukewarm, but IBM was trying to muscle into existence some absolutely dreadful ideas.

    That said, OS/2 2.1 and 3.x were godsends at the time. They did a much better job at being "better Windows than Windows" by being an incredible DOS and Windows 3.1 environment with VM memory protection. Windows 95 had a pretty UI, but architecturally it was three steps backwards (and continued the DOS hegenomy for another 5 years) and NT was an amazing resource hog.

    Everything said and done, I’m thrilled for OS X (nee NeXTStep) and Linux. M$ is no longer part of my life.

    Murray Todd Williams

  57. i don't care says:

    BSD would have replaced the linux if it wasn’t for AT&T and IBM would have been the most hated list of open source and other misc. *nix users.

  58. Tim says:

    Well, several companies were still in the running when Windows 3.0 was out. DEC had GEM, Desqview would likely have GUIfied sooner than later and OS/2 would certainly have taken off. Even Geoworks Ensemble had WYSIWYG output (that looked great on a 24-pin printer), loaded far faster than Windows, had better graphics and included some basic productivity applications. In fact, I believe that had they gotten their SDK out before Windows 3.1 was released, they might still be around, even as a niche player.

  59. David Candy says:

    Kent, I said

    Microsoft, whose first OS they sold was Unix, would be babbling about Unix’s stability, scalability. (I just love pointing out to anti MS that MS was unix company).

    And someone else also mentioned xenix.

    Anonymous Coward.

    You’ve stolen one of my user names (that I use for people who blog that I physically know). While I don’t use it on Larry’s it is mine. Don’t also steal my posts. See my post #13 here. No prizes for second.

    Slashdot. First time I’ve seen that web site (I don’t normally read anything aout computers except the MSDN library) and I can’t work it out. What a mess of a page. Bet you there is no usability testing at slashdot (or they used MSN’s testers which is the same thing).

  60. Woody says:

    In the bizarro-world of OS/2 dominance, it seems to me that we might not have seen the decline (at Microsoft Office’s hands) in market share of WordPerfect, which back in the day made the bet the farm on IBM’s offering. This would have influenced both Novell and Corel’s futures, among others, indeed, Microsoft’s biggest money maker would never have taken over without the OEMing opportunities it experienced via the dominance of Windows (particularly after Win95)

  61. John L says:

    AH but there were alternatives that existed in the WIN-DOS world.

    Quaterdeck had a thing which permitted you to run a number of DOS progs together,with backgrounding and foregrounding, along with memory management that was at least equal to MS, and it gave me more ram to play with than MS.

    There was another beautiful interface named something like QM, I think that was it, Quick Menue. That was neat, you had 50 password protected windows.

    There was another ( which may still be in existance ) which ran on very ancient hardware. The name eludes me, but the GUI was very nice.

    In other words, back then MS and OS/2 were not the only game in town. The cynic in me asks why it took until version 3 to get win going.

  62. manos says:

    well to me ,sparing all the details, the wining OS would be the one that would be user friendly. i.e. the secretary doesnot have a clue about pc’s but still can write her report to the manager. My 5years old bro can open the pc and play a game.

    Win did this and selled it good. Wining home users and make them hang upon you leaves u all the time you need to make decent serves or whatever.

  63. Bob Console says:

    What if the Atari ST had used a 286 instead of a 68000. Been MSDOS compatible as a result yet ran GEM. Also what if the Amiga has used a 286? Man the multitasking MSDOS would have been great!

  64. Anonymous says:

    » What If Windows Never Got Big?  InsideMicrosoft – part of the Blog News Channel

  65. Tim says:

    Then again, some might argue that Excel, especially, and Word, gave Microsoft the capital it needed to strongarm the OS market more than Windows 3.1. I’d have to say that Excel is the one Microsoft product that was truly innovative. It changed the face of computing, put thousands of computers on corporate desks and turned the GUI into something other than a laughingstock. Of course, the fact that we’re up to Excel 2004 (at $500 a pop) and it has barely improved really shows that Microsoft was a one-trick pony.

  66. Torn says:

    Perhpas MS would be pushing the crap out of Linux right now.

    Now if only we could get windows on the linux kernel my world would be complete 🙂

  67. Nobody says:

    I think you’re overlooking Bill’s drive to be the big dog. He wouldnt have taken second fiddle to IBM for long. My bet is that MS would have jumped on the Linux bandwagon earlier than IBM did. We’d now be lovin Bill and hatin big blue. Sounds funny, no?

  68. kstjohn says:

    As a former DesqView user I can definitively state that there was no way on earth it had anything better than Windows but for a short period. When Win3.1 came out DesqView was dead in my mind. The reality was that DesqView would crash often when running dBIII or Lotus 123.

  69. Sid Boyce says:

    Linux was bound to succeed. Initially it was designed to run on 386 machines only, people saw what it was capable of and decided they wanted it on their hardware which was as different from the 386 as you could get, consequently the number of ports snowballed and now under a common kernel source, it runs on just about any hardware platform, from embeded stuff like wrist watches right on up to top or the range IBM Z-series mainframes. Except for NetBSD, that has never been achieved as traditionally one operating system has been tied to one hardware platform. The speed, pace of development and the diversity of Linux astounded me, that’s the reason I’ve stuck with it since Linus’ first kernel was put up for ftp.

    I had to endure snyde remarks from colleagues using Windows, but I could always interoperate with them using only Linux for all work and private tasks which are as rich as anyone needs. Linux is achieving success and will grow by its performance and solid reputation and without any multi-million dollar adverts except the ones that Microsoft adorns many websites with – almost every advert for Linux comes from Microsoft.

  70. SteveOV says:

    Simple really –

    If MS had failed to get windows running in protected mode, and OS/2 took off, Microsoft would have simply stolen the codebase for OS/2 and re-badged it ‘Windows95’

    Back to square 1 with your speculations.

  71. Kenneth says:

    I remember OS/2 2.0 – we had it a year or two before Win95. That was ages after Win 3.0 – and still OS/2 was crappy for the home market. (Anyone remember the forced ‘learning to use a mouse’ as part of the installation? I still laugh today). Warp came after Win95, and never had a chance.

    Windows ruled because it appealed to the mass user market. Without Windows, Apple may have been larger, or perhaps a Microsoft OS/2 would have come later and so been better. Somebody would have filled the void, and I can’t help but think we’d be in much the same world.

    IBM and Linux still appeal to the geeks and the industrial market. Perhaps they run the machine better? Who cares? I want icons and pictures and all the other things.

    Can anyone tell me of a single IBM mass-market success? They are always there at the beginning, but seem to fail to sell.

    (The PC itself doesn’t count – the cloners marketed and sold those)

  72. Anonymous says:

    Excel is not something new. There was visicalc in Apple II. Microsoft never produced anything significant that didn’t exist before. If you think that way, it shows how much you don’t know about history of desktop applications.

  73. JD says:

    Larry,

    nice tip ‘o the hat to a great developer getting stuff done. It does go to show how the right person at the right time with the right skills, can make a big difference.

    The commenter dave leavit above had the right answer though — protected mode running of DOS programs was not unknown. DESQVIEW came to my mind as well, but there were others. There was a fertile field of offerings from many quarters, and many were judged to be quite good.

    Now that’s where the ‘right place’ comes into mind. Even if others had done it (and I don’t know the timing, but assume it was concurrent) the developer who got the work done at Microsoft and got it noticed and got it approved… that is the winning combination. Microsoft had credibility in making an environment to succeed DOS (even though Win 3.0 was by far the most unstable environment I have ever run, it was indeed a winner).

    You have to ask yourself, does the Microsoft of today have that same fire, where the big political movement toward the "one true solution" can be thwarted by a cool demo and proof of concept hack?

  74. SteveOV: Why on Earth would Microsoft have to "steal" the codebase for OS/2? Microsoft owned it outright (along with IBM).

  75. > Linux was bound to succeed.

    Once it existed, yes. But even Linus admitted early on that, if there had been a 386-running kernel he could use, he’d have used it instead of writing his own (and nobody would know the name Linus Torvalds today). He worked on his kernel because none of the alternatives (Minix, copyrighted and not open source – Coherent, closed source – BSD, not in a useable state unencumbered) was right at the time.

    The emergence of the opensource 386 unix distros was inevitable, but we have to thank AT&T for the existence of Linux. Without them, 386BSD would have probably been out earlier, and taken over the niche that Linux has today.