In the early 1990s, we had no idea where the computer industry was going, what the next generation would look like, or even what the driving factor would be. All the developers back then knew is that the operating systems available in server rooms or on desktop computers simply weren’t good enough, and that the next generation needed to be better—a lot better. This was easier said than done, but this problem for some reason seemed to rack the brains of one company more than any other: IBM. Throughout the decade, the company was associated with more overwrought thinking about operating systems than any other, with little to show for it in the end. The problem? It might have gotten caught up in kernel madness. Today’s Tedium explains IBM’s odd operating system fixation, and the belly flops it created.
I personally really loved using OS/2 over the past ten years or so. There’s something quite elegant and appealing about the operating system, and I consider it the best way to run Windows 3.x software there is – it’s entirely built-in. The world would’ve been a very different place had IBM managed to take the operating system crown for the PC industry – or the Mac, for that matter, through Talingent.
That would be _Taligent_ .
They spent billions trying to get an OS to work in the 90s. Now they’ve bought Redhat for $34 billion for a free operating system.
The business model of red hat is not about OS development, but rather support contracts. Which is what IBM bought, and it kind of fit perfectly with the current IBM structure, which is mainly a big legal firm selling licenses and support contracts. There is some HW and research here and there, but IBM today has little to do with the type of business it was 10 or more years ago. Except for the lawyers, IBM has always been big in legal and support contract revenue. They just figure out there is no point in them developing HW or system at the scale they used to, they simply can no longer compete there anymore. Not that they were ever particularly competitive, IBM coasted a very long time on the inertia of being the first big player in the computing business.
Frankly I’m surprised they have lasted as long as they did, they have had atrocious management for over 30 yrs or so. It’s a testament to how big they once were they can live off their fat for that long.
So, what were Apple’s red card teams up to?
Yes different… we have to keep in mind that, ultimately, it wasn’t quite an “odd operating system fixation” – the underlying goal of OS/2 (and other IBM OS projects) was to return to IBM the control over computer market …so perhaps it’s even better for us that it didn’t succeed.
zima,
We shouldn’t dismiss that the microsoft monopoly wasn’t such a great outcome. Microsoft deliberately set out and succeeded in killed off a lot of of market competition by abusing their monopoly market and punishing vendors who gave other platforms fair access to the market. This is a huge shroud over the PC industry as it went from numerous platforms down to just a couple viable options. Even apple were on the brink of extinction, though ironically microsoft help it survive a total financial collapse.
Who knows what could happen in an alternate timeline. IBM failed to market itself to general consumers. If the company had a good strategy to enter the household market with OS/2, I saw little evidence of it. I don’t ever remember seeing an OS/2 section in the computer stores or magazines that were big back then. It’s swell that OS/2 worked with windows applications, but IBM didn’t seem to have much success with branding itself in a windows dominated world.
It’s hard to say definitively what might have happened supposing that OS/2 did succeed. Would it automatically have become a microsoft in terms of abusing the market? Would they have killed off sun microsystems technology, the way microsoft did? Is it necessarily true a successful company will always, given the opportunity, necessarily end up becoming a monopolistic tyrant? Is there ever a time when a plethora of companies can share the market in a stable fation and bring about healthy competition for consumers? I think the answer may have been yes in the distant past, with many thousands of small mom & pop shops where owners are running shops themselves. That was how capitalism ought to work, where anyone could have a good shot at being a successful business owner if they worked hard. But these days things are very different, wall street accumulates all the wealth and ownership to no end. They’ll cross city/state/nation boundaries buying out or crushing all the small family owned independent companies in the process. They’ve systematically taken the concept of wealth through hard work and replaced it with wealth through controlling ownership. In this day and age, I think it’s very hard to fight the corruptive influences of wall street. I don’t know that it would be possible for those in power to stand up for competition. I think that without regulation (or at least the threat of regulation), competition dies completely.
“Other/numerous platforms” or “other PC OS?” There were other platforms in the past, but MS didn’t really target them (it even supported them – by providing MS BASIC for many, or even fathering some like the MSX!). That the PC won wasn’t really due to any abusive MS action; and when talking about the PC, it was always mostly about MS OSes – it never really was “numerous platforms”, at least not any more than today when we have for example (even better than in the past…) the choice of desktop Linux with its ~2% of the market, BSDs or (ducks) Haiku.
Hm, still around the year 2000, there was an OS/2 section in my local computer magazine called “Enter” (which I think doesn’t exist anymore); but then, that magazine also once included BeOS… (and how I got my copy of it in the dial-up era)
Well, in the 90s it was a different IBM than today, still largely the old one, not yet totally disconnected from heavy-handed ways of its mainframe era… (how did MS kill off Sun tech? Java or Libre Office sort of thrive…)
And as always 😛 I think you look at the past with rose-tinted glasses, big ~financiers etc. were always ultimately dominating, and capitalism even rared its ugly head more often in the past (child labour in XIX-century factories anyone? Or such easily preventable diseases like https://en.wikipedia.org/wiki/Phossy_jaw which went neglected for many years in the name of profit? (more links:
https://en.wikipedia.org/wiki/London_matchgirls_strike_of_1888
http://theconversation.com/meet-the-matchstick-women-the-hidden-victims-of-the-industrial-revolution-87453
https://www.rcseng.ac.uk/library-and-publications/library/blog/phossy-jaw-and-the-matchgirls/ ))
I really love these tech blogs where you can learn about “phossy jaw” out from an article about IBM’s attempt to dominate the OS market in the 90s. So trivial. Yet so interesting. I would gladly continue by saying that IBM hired a lot of gays, not because they took side or the cause, but because they were hard workers, not having a family to sustain, hence they were more profitable than regular straight guys. But it struck back when aids appeared, mostly because of health insurance the company was providing to its staff. The world is not driven by intelligence and good will, but by money, greed and power. We’re just more evolved monkeys after all, using silicon chips, lawyers and nuclear weapons instead of wooden stick and rocks to fight against each others. And of course we claim ourselves from God’s heritage to makes things looking good and legitimate.
Gods, plural – each larger tribe with its own magic sorcerer(s) in the sky…
(glad you liked the trivia from my comment 🙂 )
By platform I meant the operating systems more than hardware in this context. IBM dominated PC hardware in the US (especially with clones), but they practically ignored the operating system market until it was too late and they had already given microsoft the OS market on a silver platter. If it hadn’t been for that initial licensing deal, microsoft would probably not even exist today. Regardless, by the mid 80s microsoft’s monopoly tactics were in full swing.
Growing up in wintel monopoly markets made many of us unaware of the existence of alternatives. Ironically I’m still learning alot about alternate computer platforms of the past today that I didn’t know back then. The rest of the world was quite different.
There were elements of embrace/extend/extinguish. Microsoft bundled a terribly buggy and limited java implementation which fragmented java frameworks and frustrated us to no end. Back in the day I tried my hand at java development a few times, but if I wanted to be portable and target microsoft’s implementation (which I did), I had to resort to the older and slower frameworks that microsoft limited it’s implementation to. Anyways, that’s so many years ago, I have different frustrations with java under oracle today, haha.
Sure, I don’t deny that some labor practices and medical conditions have improved over the centuries and decades, we could find many more benefits too if we set out to enumerate them. Yet I’d pose the following question to you: does progress on these fronts fundamentally depend upon the widespread regression of independent local family owned businesses, which were more viable a couple generations ago? Because if not, then I don’t feel the points you mention should be a reason to disregard the merits going back to an economy based on small local businesses rather over huge wall street monopolies.
Hm, in the 2nd half of the 80s MS still worked with IBM on OS/2…
But Sun tech wasn’t “killed off” (what you wrote) …and well, in fact, it was MS Java (kinda better from user POV …the one from Sun was subpar in its ways) implementation that was killed 😛 / stopped soon enough, didn’t have that much of an impact… But I take it that you have the same issues now with Google fragmenting Java, right? Right? 😛
And well, I didn’t really imply that progress “depends upon the widespread regression of independent local family owned businesses”, just that the latter is largely an illusion (how do we count aristocracy/magnates/monarchies/nobility/serfdom/slavery of the past?)
zima,
MS was happy to give IBM that impression. It was a strategically good move for microsoft, in case OS/2 succeeded they would be part owners of OS/2 rather than competing against it. However in hindsight we know that microsoft was happy to let OS/2 fail in favor of DOS+Windows.
Mmicrosoft’s implementation was inferior. Sun’s version of the technology was evolving, but microsoft’s version bundled onto every PC held it back. Microsoft’s JVM was IE of Javas, which held back the java ecosystem as a whole. This is classic microsoft embrace/extend/extinguish strategy at work.
Do you think android & google hurt sun? Conceptually I can see how it would. Sun&java had already lost the desktop war to microsoft (leaving it with enterprise & some embedded devices) and apple users were directly prohibited from installing it on their iphones. If android/google had supported sun’s version of the technology, it could have revived sun in a big way. So while I think android could have had the potential to help sun, I don’t think it’s existence necessarily hurt them. The two biggest desktop & smartphone OS companies already had sun on it’s deathbed for consumer app markets.
@post by Alfman 2019-03-07 3:13 pm
But it wasn’t just “impression”, MS was really working with IBM on OS/2 (and really seriously on NT, which was still to become the next OS/2 version), so it’s hard to say that “by the mid 80s microsoft’s monopoly tactics were in full swing” – maybe by ~1990, with the success of Win3.x.
MS Java was better performing on hw that was common back then, that was more important on much slower PCs of the day (hm, and wasn’t it explicitly a user-initiated download, not included? I seem to remember I had to download it…)
Java creator himself said that Google has slimed Sun (and Sun/Java hardly had a chance on the desktop, no matter what MS did or didn’t do – it was after all officially supported on OSX, but even there devs chose Obj-C/Cocoa; so why would they choose Java over more native development on much larger Windows platform? We all quickly moved from Azureus to uTorrent 😛 ). Android Java is stuck at 7 IIRC, that done more to fragmenting it than MS ever did… (hm, Google took existing Java ecosystem, molded it and took into their own direction with its mobile OS, and now that it succeeded and gathered momentum looks like Java might be deemphasized in favour of Google’s own Dart/Flutter; that’s not “embrace/extend/extinguish strategy at work”? 😛 ) Supposedly Sun was by then in too fragile financial position to enter into legal dispute…
PS.
Every wall street monopoly was once a small local business. Thing is, most small businesses want to grow. Few succeed…
zima,
I happen to disagree. I really think everything happened the way microsoft wanted it to happen and IBM was outplayed because microsoft successfully managed to put IBM’s guard down. Partnering with IBM was a better strategy than competing with them outright. I’m suggesting that the deception and manipulation traits that you believe were adopted later on were actually an integral part of microsoft’s strategy from the start. In other words, their actions were calculated, it wasn’t just blind luck. Microsoft wouldn’t have become a monopoly if they didn’t play IBM the way they did, I don’t think that’s in dispute. What we don’t know is what the computer market would be like if not for microsoft, we can only speculate. While my guess isn’t better than yours, I’d argue that the very same playbook that microsoft used to beat IBM would ultimately turn out to be more harmful to competitors in the computer market than IBM acting by itself.
You have made the point that maybe there would have been another abusive monopoly if it weren’t for microsoft, but I still think the computer market could have been more competitive if microsoft hadn’t gotten such an early monopoly from IBM.
Sure, both google and apple hurt sun’s mobile prospects. Google by building a competing clone, and apple by banning it outright. However on the desktop microsoft causes a significant amount of damage, which I can attest to by having been a java developer at the time. The java experience for developers & end users was worse off because of microsoft’s buggy outdated JVM, which was a stalling strategy until they could release their own competing .net application platform.
Not always, some startups start up with way more money than others. But that wasn’t even my point. When mom & pop shops were the norm, a middle class family had a real shot at opening up their own local mom & pop shop and supporting themselves. That kind of opportunity is slipping away for most of the middle class for several reasons. The costs to buy into a new business are significantly higher than they were for previous generations. The costs and fees for commercial property are astronomical, for those of us who live in areas where it’s against codes to open a commercial business at home, it’s a non-starter. You’ll say it was always difficult, which is true, but it’s worse now. The real kicker is that your business is not competing with other mom & pop shops as in the past, nowadays you enter the market competing against well established multinational corporations formed through years of consolidation and exponential growth. It’s like entering a game of monopoly mid-game where the other players already own everything.
So it’s not that your wrong about anything specific, but the significant decline in the viability of mom & pop stores is pretty well established. We cannot deny that corporations have consolidated and grown substantially in the past several decades and we also cannot deny that their growth has come at the expense of small local businesses. I guess some people are ok with this, but I think it’s hard to make an economic case that this wealth accumulated at the top hasn’t hurt the lower & middle classes.
@postby Alfman 2019-03-10 12:04 am
Sure, IBM guard was down – but OTOH that’s kinda another argument against “by the mid 80s microsoft’s monopoly tactics were in full swing” – if they would be, IBM would most likely take notice… Just executing a very succesfull strategy is not abusive. And, again, MS even worked thurough the 80s on competition to PCs; seems they simply didn’t miss any chances, that’s beeing savvy.
Well, as an end user with paltry 64 MiB RAM at the time I happened to appreciate MS Java, it ran better. 😛
And don’t forget that your small mom & pop businesses from the past were largely unable to tackle big problems of today; they operated on a local services level.
Several times in the article was mentioned a goal to replace specialized solutions (OSes and kernels) with one ultimate generic one that fits all sizes.
Over the years of working at various IT projects, I’ve seen such things many times and practically always they fail. It is common part of a person to try to unify specific seaprate systems to singe common one. But as mentioned also in the article, the single generic solution rarely if ever can match the requirements of all customers. So you may even create such a generic platform but to actually deploy and use it, it usually has to be customized, if not hacked to get to work within requirements. So you may have a generic platform but every solution would be unique and custom.
Or the other way – people set the goal to create one unified platform but the requirements and the scope grows out of proportions and to deliver something within given time and budget scale, they opt to create a specialized focused solution. One delivered, they plan to return to the dream of unified platform but again being faced with constraints, opt for a specialized solution.
To me the specialized and focused solution approach is totally acceptable. Do it couple of times until common patterns are emerging. Unify those and now use the common part in the future projects, Go back, search again for common patterns, unify those and ultimately you end up with some sort of a library or toolset or just a common layer that is shared among all specific solutions. However I personally do not believe in a possibility to have one product that can fit all sizes.
Even Microsoft which probably has one quite large “platform” called Windows, do have quite a few parts in common between desktop, server, mobile XBox, Hololens etc. Yes the kernel probably is but the rest is customized for specific needs. And then again it has taken lots of iterations and trials before Microsoft managed to filter out the common parts across all the target platforms to end up in unified kernel. So they made special solutions first and then generalized and ended up with common kernel. Not the other way around. Specific to common way usually works, opposite way I don’t think so much.
> Even Microsoft which probably has one quite large “platform” called Windows, do have quite a few parts in common between desktop, server, mobile XBox, Hololens etc. Yes the kernel probably is but the rest is customized for specific needs. And then again it has taken lots of iterations and trials before Microsoft managed to filter out the common parts across all the target platforms to end up in unified kernel.
While your argument is valid, I’d like to point out that there’s a whole lot more than you probably realize that’s shared between Windows platforms. XB1, for example, actually uses an essentially unmodified version of Hyper-V under the hood to run games (they get a dedicated VM that looks like bare metal with a thin veneer on top, that’s part of why it’s been so easy for Microsoft to get classic XBox game support on the XB1). Much of userspace below the level of the UI shell is also almost identical on pretty much all modern Windows platforms except IoT, Edge code powers web usage everywhere except servers (which for some reason still only offer IE 11), DirectX is present with identical API’s pretty much everywhere that uses it, UWP support is also largely unmodified on all platforms, NTFS is the filesystem of choice on pretty much all the platforms, etc.
Put differently, the high-level differences between Windows ‘platforms’ are more akin to the differences between variants of Ubuntu than the differences between disparate Linux distributions like NixOS and Alpine.
Out-of-box experience:
OS/2 Warp 3: Sound didn’t work. Only VGA video. CD not recognized. The Internet did not work. Yes, after much twiddling and fiddling I was able to get all of the above working, Hours and hours of fiddling.
Windows 95: Sound, SVGA and XGA video, CD supported, Internet connection worked.
OS/2 was (and still is) amazing…for an engineer.
So if IBM makes chips and operating systems, why couldn’t they just design a chip to mitigate the performance loss of a microkernel by designing circuits that accelerate microkernel specific code?
dark2,
The main reasons why microkernels, especially back then, were slow are not microkernel specific. Microkernels just push more stuff out of kernel space, and forces those subsystems to communicate via messages.
Any hypothetical circuit that makes interprocess communication faster will also make monolithic kernels seem faster. And people will still notice the performance difference and choose the faster one.
That would have been prohibitively expensive. Plus there is nothing intrinsically special about microkerneles from an architectural standpoint. Also OS/2 was not a microkernel as far as I know.
Hi there.
The OS/2 users still hangs out at OS2World.com and some other sites. You are all welcome to install OS/2 on a Virtual Machine and join us !!!
Regards