Larry Osterman's WebLog

Confessions of an Old Fogey
Blog - Title

Tipping Points

Tipping Points

  • Comments 84

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.

  • 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.
  • 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 :)
  • From the dark side. . .

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

    Each "what if" root has many branches
  • 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 :)
  • 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.

  • 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.
  • 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?
  • 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!

  • 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.
  • 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?" :)
  • 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 :)
  • 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)
  • 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.
  • 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
  • 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.
Page 1 of 6 (84 items) 12345»