Slashdot Mirror


The Sad Parable of OS/2

Still-in-Mourning writes "IBM's first 32-bit version of its advanced PC operating system was released 10 years ago this month. It was better than anything around, yet it failed. Its hopes were pinned on many of the same things we hope today will bring Linux to the forefront. What lessons are to be learned? Will we learn them? A glimpse of a sorry chapter in computing history."

19 of 550 comments (clear)

  1. IBM killed OS/2 by Eric+Green · · Score: 4, Informative
    Remember, OS/2 was originally released as part of IBM's PS/2 attempt to re-hijack the personal computer industry. The personal computer industry wasn't buying it -- they had no desire to put themselves back into thrall to IBM.

    It's hard to believe, in today's day and age when Microsoft is the "evil empire", that there was once a day when Microsoft was the scrappy upstart and IBM was the "evil empire", but that's what the situation was like for most of the 1980's. In the end it did not matter how good OS/2 became... nobody was going to put their company at the mercy of IBM again.

    By the time OS/2 Warp (32-bit OS/2) came out, if you mentioned OS/2 to anybody in the computer industry, they'd say something like "You mean that runs on something other than IBM PS/2 computers?". Unlike what somebody else here mentioned, everybody in the computer industry knew what OS/2 was and what it was capable of doing. But a) they didn't know it ran on anything other than IBM equipment, and b) they weren't interested in putting themselves back into thrall to IBM again.

    In the end, politics, not technology, doomed OS/2. The politics of Linux are completely different from the politics that doomed OS/2, and I can't think of any lesson from the OS/2 saga that applies to Linux.

    -E

    --
    Send mail here if you want to reach me.
    1. Re:IBM killed OS/2 by Eric+Green · · Score: 3, Informative
      The PS/2 was a computer that IBM released in, I think 1986. Compaq and a number of other companies had come out with wildly successful clones of the original IBM PC, and IBM realized that they'd given away the personal computer market. So they created a new computer bus -- the MicroChannel Bus -- incompatible with the bus in the original IBM PC (the bus that Compaq and others used). They created a new operating system -- OS/2 -- in conjunction with Microsoft, that ran only on their PS/2 (Personal System/2). Then they dropped all their "old" PC-compatible machines, and you could only buy a PS/2-compatible machine from IBM. They felt that business would buy PS/2 machines from IBM because business bought IBM, and they would not license the patents to their Microchannel bus to other personal computer vendors, so they would have control of the personal computer market once again.

      But it didn't work like IBM planned. It was an unmitigated disaster. IBM sold only a few thousand machines, and had been geared up to sell millions. 16-bit OS/2 on a 16 mhz 80286 microprocessor took a half hour to boot, and there were no expansion cards for the new 16-bit MicroChannel Bus. They swiftly rushed their old "PC-compatible" machines back into production (calling them the PS/1 and other names like that to imply that they were only half as good as their PS/2 machines), but the damage was done -- IBM was never again the #1 maker of personal computers. The PS/2 lingered on for another couple of years as IBM continued to try to push it, and was mercifully put out of its misery when the industry migrated from the 80286 (16-bit) processor to the 80386 (32-bit) processor.

      Whenever you think about the eventual fate of OS/2, you have to recall how it originated -- and what IBM was trying to do when it created OS/2 in the first place.

      -E

      --
      Send mail here if you want to reach me.
    2. Re:IBM killed OS/2 by jht · · Score: 3, Informative

      It wasn't quite as bad a disaster as that - PS/2 sold fairly well, but not well enough to kill the cloners. And there was a 386 in the line from the get-go, though ALR and Compaq had released the first 386 PC's a few months prior. The actual release of the PS/2 was in early 1987. Micro Channel was a much more advanced bus than ISA represented - mind you, this was before the whole industry coalesced around the (not yet invented) PCI bus. Feeling the pressure from IBM, the rest of the industry got together and devised EISA (extending ISA to 32-bit goodness and backwards-compatibility) - a decent Micro Channel competitor that carried the clone market for a few years in the interim before Intel pushed PCI out.

      The goal was definitely to lock up a new standard, though. At first. IBM offered to license Micro Channel, but at very high royalty rates that effectively left no room for competitiors. OS/2 started out as a vaporware project that relied heavily on Microsoft to manage big chunks of it, and ultimately became IBM's flagship OS and their "open" competition to a rising Microsoft. Windows 3.0, OTOH, started out as a way for Microsoft to hedge their bets against slow adoption of OS/2 - after the first couple of years IBM had opened up to the reality that they needed to support the cloners, too. When Windows took off and the big MS/IBM split happened, Microsoft got to keep the OS/2 3.0 project that was being planned at that point. IBM decided their future was in porting OS/2 to their new Power series chips. Which ultimately fizzled out.

      The Microsoft part of the project became Windows NT. OS/2 itself (Warp was a marketroid decision to add the codename to the product) had wonderful Win16 capabilities back in the Windows 3.x days - but Windows 95 came out conveniently after IBM's license to Windows source expired and that was the commercial death of OS/2.

      I think the last PS/2 was canned around 1995 or so, maybe a hair later. There were some good products made for the MCA bus, mostly connectivity products. It was a far better bus than ISA, but the market (and IBM) killed it easily.

      The legacy that PS/2 left us in the end was mainly the mini-DIN connectors for keyboards and mice. IBM sold a decent number, but not enough to justify a separate line of PC from the mainstream. Apple's really the only folks who have ever pulled off a different standard over the long term.

      (This is also a good argument as to why Apple should never go to Intel as chip vendor - IBM had a good alternative OS, a neat box, and a better mousetrap, but couldn't differentiate themselves enough to thrive.)

      I may be slightly off on a detail or two, but I think my recollection is fairly clear on this. Feel free to correct specifics, folks!

      --
      -- Josh Turiel
      "2. Do not eat iPod Shuffle."
  2. Did you know who wrote OS/2??? by Captain_Frisk · · Score: 5, Informative
    I'll give you a hint, it wasn't IBM

    Actually, it was (gasp) MICROSOFT (gasp). Think about that before you flame!

    Here are googles top 2 links with more information.

    1. Link
    2. Link

      and the google search itself
    1. Re:Did you know who wrote OS/2??? by os2fan · · Score: 5, Informative
      Microsoft wrote some, IBM wrote some.

      Likewise, you can say IBM wrote Windows.

      The really good bits (REXX, IPF, WPS, PM, IFS, Program Manager, File Manager) are IBM stuff. The bad bits (the DOS coffin, 16-bit stuff) are Microsoft's stuff.

      IFS forst appeared in the DOS world in PCDOS 4.0. IBM wrote that.

      IBM had virtual machines before Microsoft *existed*. File and Program Manager appeared in OS/2 1.1 or 1.2. Microsoft borrowed these for the Windows 3.x shell apps.

      REXX and IPF are IBM mainframe stuff, using standard bits in different operating systems.

      WPS is IBM's invention: the shell, and even the colours were borrowed by Microsoft. The teal background first appeared in OS/2 2.11, way before Windows.

      And more, IBM tried to support existing machines, and not only the latest and greatest. IE they support the idea of using your OS on an old machine.

      --
      OS/2 - because choice is a terrible thing to waste.
  3. Re:I don't get it. by Anonymous Coward · · Score: 2, Informative

    What exactly in the in the 286 architecture prevents the use of a multitasking operating system?

    The lack of memory protection. The x86 line didn't have an MMU that could be configured to protect apps from each other until the 386.

  4. Re:OS/2 Still In Use.. . by neurojab · · Score: 5, Informative

    It's a little known fact that many ATM machines use OS/2... even the new ones. That means millions of people use OS/2 every day and don't even know it. The funny thing is that they WOULD know it if they used an M$ OS. How would you like the "blue screen of death" when you're in the middle of a transaction?

  5. Re:OS/2 by Anonymous Coward · · Score: 0, Informative

    I did, but he says he preferred 5 year old boys more and was soon arrested for sodomizing his own nephew. Sick bastard. But hey, that's what you get from those freaks at Berkeley.

  6. Re:OS/2 Still In Use.. . by Anonymous Coward · · Score: 1, Informative

    In the Netherlands Windows-NT is being used for ATM's.... and it shows!
    Though relatively stable, I've seen 'em crash causing the Windows-NT splash screen to be shown and rendering the ATM inoperable.

  7. Re:Win95 didnt kill OS2, Microsoft did. by stubear · · Score: 3, Informative

    exclusive EM contracts aren't illegal until you're ruled a monopoly. When Microsoft established these contracts they were not a monopoly. Sure, over time they became one because of the exclusive contracts.

    Dell tried to sell Linux workstations. Their endeavors failed and they dropped the program altogether. They still sell servers with Linux preinstalled but that's it.

    Sun is failing as a hardware/os/software company due to Linux. SGI isn't in much better shape though they have one of the most lucrative industries in the world clammoring for their machines - Hollywood. Apple is doing well though they are having a tough time climbing out of their niche market. Be, well, be was, and won;t be anymore.

  8. Re:I don't get it. by steveha · · Score: 5, Informative

    What exactly in the in the 286 architecture prevents the use of a multitasking operating system?

    That was not the problem. The problem was writing a multiasking operating system that would run all the DOS apps (which were important at the time).

    When the 286 was in protect mode, some of the instructions worked differently than when it was in "real" mode (8086 compatibility mode). Result: you could not execute DOS apps; they wouldn't work.

    So, how about making a DOS virtual machine? Well, the 386 has features that make it easy to spin up multiple real mode virtual machines, but the 286 didn't have those features. A purely software virtual machine would be very slow.

    So, how about switching out of protect mode and running real mode code in the 286's real mode? That was the only option, so Microsoft took it. However, Intel had not designed the 286 to do this. There was an instruction to start up protect mode, but no instruction to leave it and go back to real mode! Microsoft wound up programming the keyboard controller chip to actually reset the CPU, many times per second, to switch to real mode.

    Because DOS apps ran in real mode, they owned the whole machine: all memory, all devices, etc. So if a DOS app crashed, it would take the whole machine down with it; a crashing DOS app could trash OS/2, and there was no way to prevent it.

    Even worse, the 286 did not have features that would let you virtualize the hardware, and DOS apps liked to talk directly to the hardware. All DOS apps liked to write directly to the video card, rather than going through the BIOS, and the 286 didn't really help you solve that problem.

    So the OS/2 1.x "compatibility box" could only run a single DOS app at a time.

    Meanwhile, Microsoft sold Xenix 286, which worked perfectly well. Alas your Xenix 286 programs either had to be less than 64KB each, or else they had to deal with near/far pointers (yuck), but Xenix 286 worked. Microsoft never tried to do a GUI desktop for Xenix, but it would have been possible.

    It appears to me that the article writer is trying to excuse Microsoft's lack of skill by pretending that the task was impossible.

    No, it really was impossible to write an OS that would run decently fast on the 286 hardware of the day, would multiask old DOS apps, and would be reliable. The 286 was just too broken.

    steveha

    --
    lf(1): it's like ls(1) but sorts filenames by extension, tersely
  9. Re:Windows 95 applications killed OS/2 by sconeu · · Score: 4, Informative

    Heh, and now that MS has a stable OS, the apps have all gone down the shitter.

    What apps? Just about every commercial application on the market five years ago has been replaced by a Microsoft clone.


    He's right, and the exception that proves the rule is Quicken. The only reason Quicken still exists is that the FTC (for reasons that are still unknown, given how merger-happy it seemed then, and still does) nixed the MS buyout of Intuit.

    --
    General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
  10. There was plenty wrong with OS/2 2.x by lseltzer · · Score: 2, Informative

    OS/2 2.x was a great piece of software in many ways, but IBM missed the mark on several important points, some of which the linked story gets.

    For example, it should be perfectly obvious by now, and was to many even at the time, that preloads were a critical factor. Microsoft appreciated this much earlier than IBM, who couldn't even convince their own PC company to preload OS/2. Remember that OS/2 2.0 came out long before Windows 95 - If OS/2 really was a better Windows than Windows, like IBM claimed, no threat Microsoft could make should matter. The IBM PC Co should have been happy to preload OS/2 2.0 and dump Windows 3.x. (Remember IBM even had their own DOS on the same code base, they didn't need Microsoft at all.) The fact that IBM PC Co could never be convinced to preload OS/2 is damnimg evidence that it was never all it was cracked up to be.

    There were dozens of important problems. Among which:

    • 32-bit OS/2 driver support lagged badly behind Windows driver support, which was a far higher priority for every device vendor out there. For the first year or so there were maybe two graphics cards with OS/2 drivers shipping, which meant that you were stuck with standard VGA on any other system (I don't think they had a standard SVGA 8x6 then).
    • The minimum RAM requirement for OS/2 was 4MB, a high-end requirement at the time. By the time Windows 95 came out in 8/95 4MB was mainstream, but it was a real problem for IBM, especially since they were trying to sell into the existing installed base.
    • The OS/2 Netware driver shipped very late and was a steaming turd when it did ship. Good Netware support was about as critical as can be, so it's hard to see how IBM didn't place a higher priority on this. IBM's own LAN Manager networking was rock solid once you had it set up properly, but as others have pointed out, it was a bitch to set up.
    • This is a matter of taste, but I always thought the Workplace Shell was an awful user interface. The context menus were crowded with complicated and irrelevant distractions. You really needed to use HPFS for your file system because the disconnect between the naming system in OS/2's FAT and the WPS was complete. Fortunately, HPFS was a good file system for its time (invented at Microsoft by the great Gordon Letwin), but if you wanted to dual-boot from OS/2 to DOS/Windows you were screwed.

    Microsoft was hardly friendly to OS/2 after it washed its hands of it in the 1.3 days, but all the big reasons for its failure in the market have IBM's fingerprints all over it. Them and a gang of fanatics that make the worse Linux advocated look downright boring.

  11. 1.0 vs. 2.0 by Eric+Green · · Score: 5, Informative
    2.0 was designed from the get-go to run on pretty much any 32-bit hardware out there. IBM had abandoned the notion of trying to hijack the personal computer industry by that time. The problem is that by the time it came out, everybody in the computer industry was operating under the notion that OS/2 was for the PS/2. Which was true, in the beginning.

    Regarding IBM and Microsoft and OS/2, I've read some reminiscing by one of the industry pundits who was there at the meeting where IBM blew off Microsoft. Bill Gates showed up with all these charts showing Windows as a little side project on top of IBM/Microsoft OS/2, and IBM blew him off. Yep, that's right, IBM blew off Microsoft -- NOT the other way around. That was apparently when Bill decided that Windows was going to be a totally seperate operating system not reliant upon anything IBM (Chairman Bill does NOT like being blown off by arrogant IBM execs!), and that was when Bill decided he was going to borrow some tactics out of the IBM monopoly handbook, such as bundling, "vaporware", and per-CPU pricing.

    Now, I'm not going to argue about whether the Microsoft monopoly on personal computer desktops is good or bad. I'll just point out that an OS/2 monopoly would probably have been even worse -- because IBM is a hardware company as well as a software company, and undoubtedly would have used their hardware muscle to squeeze out the kind of white box clone business that kept Linux alive for many years before the major vendors discovered Linux.

    -E

    --
    Send mail here if you want to reach me.
  12. EComStation by Anonymous Coward · · Score: 1, Informative

    The latest release of OS/2 is by Serenity Systems.

    Take a look at www.ecomstation.com.

  13. Give Wine a chance by marm · · Score: 3, Informative

    Wine is considerably more, err, byzantine.

    Hardly surprising given what the OS/2 Win16 compatibility layer was: Windows 3.1 run in a virtual 286 (thus in standard, not enhanced mode, which is why some apps wouldn't run), without the Program Manager. IIRC there were two versions of OS/2 you could buy at one point, one with a cut-down copy of Win3.1 included (that Microsoft let them include it was a legacy from agreements signed during the Microsoft/IBM co-operation days, although Microsoft was still getting a licence fee from it) and a cheaper version which asked you for your Win3.1 disks during installation.

    A much more fair comparison would be with Win4Lin, which attempts to do much the same thing with Windows 98, i.e. run it in a virtual machine without Explorer, and display the application windows on an X desktop.

    Win4Lin is actually a more impressive achievement, as in order to run Win98 it has to virtualize a 386 in protected mode, which is hard, as opposed to virtualizing a 286, which is really really easy (the 386 and above has hardware especially designed to do this).

    Either way, whether you run Win16 apps in OS/2 or anything Win98 can run in Win4Lin, you have paid the Microsoft tax and are running Microsoft code.

    The other approach to running Windows software on other systems is to reimplement the Windows API. Some projects that do this are basically ports of Microsoft code (like Mainsoft's MainWin, which is used in the HP-UX and Solaris versions of IE). Just two projects have ever done this without using any Microsoft code: Wabi and Wine.

    Wabi was very successful in its time, providing a complete Win16 layer on lots of UNIXes, and something that even Wine doesn't do: an i386 emulator for people who want to run Windows apps on non-IA32 architectures. Unfortunately it never got much Win32 support and, being a proprietary product, died a death a few years ago.

    So we're left with Wine, the most ambitious Windows emulation project of them all: efficiently reimplementing all of the Win16 and Win32 APIs and ABIs, without any Microsoft code, and all as free software.

    it seems like the developers are more interested in using the code for proprietary emulation for running specific programs (games, plugins) or porting (corel stuff, etc.) than producing a general, Free, universal windows emulator.

    They are effectively rewriting about 30% of Windows, with only Microsoft's published documentation and reverse engineering as references. They have to be bug-for-bug compatible (this is the real killer). The core team is absolutely tiny compared to Microsoft's Windows development group.

    Is it any surprise that they are trying to do what they can as they get things working? They are doing stuff that people would have thought near-impossible just a couple of years ago, even by an extremely well-funded corporate behemoth like Sun or IBM.

    It would seem that Wine is the most underappreciated of all the major free software projects out there, which is such a shame given its promise.

    Give Wine some time. I know it's been a long wait already, but the pieces are falling into place right now, and it shouldn't be too long (measured in Wine time, of course ;) before Wine gets to version 1.0. When that happens, expect repercussions for years to come.

  14. OS/2 Screenshots by searleb · · Score: 4, Informative

    In case you have never used OS/2 and you are interested in what it looked like (as I was), this essay is chock full of screenshots.

  15. DIVE = WinG aka DirectDraw by operagost · · Score: 3, Informative

    DIVE stands for DIrect Video Extensions. It's not running DOS games in a window. What IBM had done was allow OS/2 programmers quick access to the display adapter instead of having to use the slow GDI, just like WinG would do later. You see, when you don't want your programs to crash your spiffy protected-mode OS, you have to virtualize your hardware.

    --

    Gamingmuseum.com: Give your 3D accelerator a rest.
  16. Re:OS/2 by j_w_d · · Score: 3, Informative

    I don't think it was the greatest. The user interface was locked in. Much like NT (microsoft). So I never really cared for it.

    You plainly never used OS/2 enough to know it. Presentation Manager (PM) was partially object oriented, while the WPS (work place shell) was fully OO. Companies like Stardock produced terrific desktops that were nothing like the default PM or WPS layouts. One of the cool things about Stardock's products was the addition of properties to file types based on inheritance, such as the property that text files were inherently editable. Click on a text file and it came up in an editor. Stardock's interface for OS/2 is very similar to KDE or GNOME, with multiple desktops, and a small windowed desktop selector. Windows 95 actually copied the OS/2 WPS or PM interface, which was quite clean and easy to use. OS/2 WARP was also the first desktop OS, besides Linux (and at that time it was not automated in Linux as it was in OS/2 WARP), to come with built-in communications (beyond some communications program such as ProComm) and internet connectivity, while Gates was still pushing that glorified BBS from Microsoft. At the time you still had to download winsock utilities from MS and install them if you wanted to use the internet. Micorsoft's prucahse of the parent program that became Internet Exporer was driven by the need to answer OS/2 and provide internet services. WAIS, Gopher, FTP and Web connections in OS/2 could be dropped as icons on the desktop. Click on them and the modem would automatically dial and connect. OS/2 was not as stable as Linux, but even when it went down, it came back with less trouble than Windows, and you could easily back step to a previous configuration, if a program installation clobbered the system with an incompatible driver or something. There was also never any necessity to reinstall OS/2 as the installation aged. This is still a common occurence even with modern versions of Windows when the registry becomes so clogged with crap the system becomes inherently unstable. OS/2's configuration files were simple, text based, and easy to fix with an editor. There is still a lot to like about it.

    --
    ------ The only greater hazard to your liberty than n politicians is n+1 politicians.