March 17, 2003

Imminent Death of UNIX and Apple Postponed. More at 11

For more than a decade now--ever since Intel's chips began to eat the SPARCs for lunch--I have been expecting the imminent death (any year now) of UNIX and of Apple, any year now. What UNIX had going for it was that it ran on workstations, but then Intel began making workstation-class chips cheap. What Apple had going for it was a more coherent and less brain-taxing illusion of what was going on inside the computer, but certainly Windows 95 was good enough.

What Microsoft had going for it was mammoth economies of scale. Selling their operating system cheap had gained them an enormous market share. This meant that development costs could be amortized across an enormous user base--and thus per-user development costs were tiny. More important, if you were an outside developer writing programs, you could write for Windows and have a large potential market, write for UNIX and have a small potential market consisting of early adopters and heavy users, or write for Apple and have a small potential market of insane cultists. It seemed to me clear that developers would gravitate toward Windows. That more on more interesting uses would appear on Windows first and on UNIX and Macintosh later (if at all). That the gap in modernity between Windows and other platforms would grow. And the other platforms would die.

That hasn't happened.

Now comes Joe Brockmeier, opining (based on a survey by Evans Data) that Windows has lost the race for developer attention:


Open source, linux, microsoft, perl, BSD, GPL, PHP, Apache, MySQL, GCC. Joe 'Zonker' Brockmeier - Corante: ...Evans Data did a little research, and the findings suggest that Linux is far more widespread than you might think: "This year, Windows commands attention of 50 percent of the developers. Roughly 40 percent focus primarily on Linux. These priorities will switch places almost number-for-number next year..."


Clearly it's time to mark my beliefs to market on this. Where did I go wrong? Where was the flaw in my chain of reasoning?

Posted by DeLong at March 17, 2003 10:00 AM | TrackBack

Comments

Well, I don't know all the reasons why, but I can give you some of them:

1. Most developers don't work on commercial for-sale applications, but rather in custom development shops cranking out corporate applications. Therefore desktop market share is less important because the app will never be sold.

2. Related to that has been the growth in web-based apps. Most web servers use Apache and rn on Linux boxes.

3. Many of the best developers learn their skills as poor student types. Linux is free. MS is not. Therefore kids coming out of school have a lot more experience developing on that platform, and it's easier for IT shops to cater to their existing knowledge than to retrain them on unfamiliar tools.

Posted by: Kevin Brennan on March 17, 2003 10:25 AM

Linux has piggy-backed on Microsoft's economy of scale, since Microsoft's success made Intel-based PCs a mass market. Linux wouldn't exist if Microsoft's success had not made PCs affordable.

Posted by: Scott Hanson on March 17, 2003 10:35 AM

The "Evans Data Corporation survey" is available in summary here:
http://www.evansdata.com/n2/surveys/linux_toc_03_1.shtml

Their methodology is not described in the sample pages, but they hired Nicholas Petreley to help them:
http://www.linuxworld.com/2003/0314.petreley.html

which is kind of like hiring Guy Kawasaki to do a study on Apple market share (or Molly Ivins to do a study on Democratic voter share, if you prefer).
Note to Kevin Brennan: I am familiar with the C.S. curricula at one community college and one state university in Northern California. The former uses exclusively Windows, the latter mostly.

I write all of this as a former Mac zealot turned Linux developer...

Posted by: Sam Penrose on March 17, 2003 10:56 AM

Scott:

YOur comment is true so far as it goes, but what does that have to do with the choice people are making to use Linux?

(It's also incomplete; the decision by IBM to use off-the-shelf parts for its PCs, combined with Intel and Microsoft's forsight in negotiating non-exclusive deals with IBM was the key to driving down computer prices by making PCs a commodity item).

Really, MS's dominance in the PC market is the only reason it has ANY penetration amongst developers. Like I said, most developers work on custom software for major corporations. Only a small fraction of them work on commercial apps and an even smaller number of those work on apps that are likely to be sold to individual users. (Anecdotally, based on the people I know and on job listings over the years, I would guess that 95% of active developers work for the corporate market).

So really, it doesn't matter what people use on their desktops. What matters is which platform the server that runs the main app is using. The last report I can find on this is from 2001, but it shows that MS has about . That woud be more or less consistent with what I said above.

Now, my personal experience doesn't agree with the report, as every app I've been involved with has been developed with Windows smewhere in the loop. So much for anecdotal evidence...

Posted by: Kevin Brennan on March 17, 2003 11:05 AM

I think there are a few issues involved.

First, more and more development is occurring in the network sphere as corporations take their aging stand-alone applications and migrate them to web-based interfaces.

In addition, a larger percentage of new development projects that previously would have been developed as stand-alone tools are being reconceived as full-scale networked applications.

Finally, many new standalone applications now have an online aspect requiring a backend server system.

With these fundamental changes towards network-based development comes the high probability that Linux is being used due to its speed, cost, security and stability. For example, the PC and console game market is taking off and many new games are either completely "online" or have an online-play option. It's a good bet that these game companies are using Linux on the back-end to guarantee server uptime and security, even though none of the front-ends are based on Linux.

Posted by: Chris Nelson on March 17, 2003 11:07 AM

Microsoft is not a great innovator. Almost everything they have done has been an imitation of technology originally produced by other entities. MS excels at reverse engineering and avoiding expensive development costs.

The original DOS was an imitation of the CP/M operating system. The developers of CP/M did not get on with IBM. IBM looked for an alternative, went with DOS and MS had its marketshare. Windows was a complete ripoff of the MacIntosh operating system. The idiots running Apple at the time were marketing types who believed they were running a hardware company, when Apple was in reality a software company. So they let MS have the Windows technology without a fight. The MS Explorer Browser is a ripoff of Netscape. On and on it goes. Microsoft is closed source software, so it is less friendly to product development. Plus, those who develop killer apps for the Windows OS will be pressured to sell out to MS or face competition from an MS clone. While this strategy has given MS a huge market share, it also chokes off much of the innovation that could be done. Thus, MS has a huge market share, but their dependence on others to develop new ideas is the reason why Linux and Apple are still in business. MS also owns a sizable portion of Apple.

Posted by: bakho on March 17, 2003 11:14 AM

economics is the answer of course!

actually as a former developer and a current financial type I can give you my 2 cents.

I developed applications on UNIX platforms because the damm thing never crashes. Of course you could screw up in your programming and cause a box to go down, but you are spared from the MS blue screen of death among other things.

Now with Linux (which is UNIX) you get a few things.

1.) free, mature, UNIX operating system
2.) cross-platform chip support
3.) ability to customize/ tune the operating system to your program's need
4.) great support via peer groups
5.) access to tons of free programs that let you do everything that MS wants to charge you for

now as a finance guy I would see the relative advantages over a MS platform as:

1.) free, mature operating system
2.) cross platform chip support
3.) ability to customize / tune OS
4.) free support
5.) did I mention it is free?

Key point ahead:
and if you are doing a web application the desktop doesn't matter (thanks go to Mosaic/ Netscape team - this is why MS needed to put these guys out of business)

plus if you are a big, slightly scared, cover your ass corporate type IBM is now fully in Linux's corner. Call it the revenge of Big Blue - they can charge huge sums for their consulting services and put all their progress into something that gives them independence over those guys in Redmond.

so what can MS do?

1.) co-opt HTML standards so internet explorer is the default platform and only works well on windows boxes
2.) keep all Office documents in some sort of proprietary format so there is a extremely high switching cost associated with free/ other programs
3.) lock in customers via long term licensing agreements
4.) figure out a way to co-opt Linux (difficult to do b/c of the Linux license - thanks to GNU)

the real key (in my opinion) was the independance and utility of the web browser.

Linux frees you from lock-in.

And it is free.

Posted by: jjj on March 17, 2003 11:19 AM

If I was a developer (which I'm not) I'd be pretty leery of writing for Windows machines-- a large percentage are standardized 'box on the desk' computers that don't give the user any leeway about what software to use. In contrast, UNIX and Apple customers are a lot more likely to poke around with their software

Posted by: Matt on March 17, 2003 11:26 AM

I think the "free" aspect of Linux is being overhyped. What was that saying I heard recently? "Linux is free if you value your time at nothing."

Posted by: Chris Nelson on March 17, 2003 11:44 AM

Scott:

And as I mentioned above, while most new developers are learning their skills, they DO IN FACT value their time as worth nothing.

The academic release of Visual Studio is several hundred bucks, plus you have to get a top-notch PC (running Win NT Server). Or you can learn to develop on Linux for free.

As a student I had about $50 a week disposable income. If I'd been a dveloper type, you can bet that I'd have learned on Linux.

Posted by: Kevin Brennan on March 17, 2003 12:13 PM

There are good reasons for Linux to be winning the server market, but client-side?! Let's face it, you don't get good GUI in HTML, nor in Java.

I guess I got spoiled by Lightspeed Pascal (Mac OS 6, circa 1986, $49 legitimate copy), but to the extent Linux has any RAD tools, they're a ripoff of Visual Studio and not the other way around. If Apple can make real OS X RAD tools that work on the underlying Unix, they'll sweep a huge share of development.

Posted by: Andrew Lazarus on March 17, 2003 12:19 PM

Kevin,

I wasn't trying to give a complete answer. I guess I was thinking of the early history of Linux, and how it got established at all in light of the Microsoft monopoly. And since I work as a BSD/Linux sysadmin for a internet software developer, I'm certainly very happy for the success of Linux, since my career depends on it :-).

We make web banking and brokerage server apps for Solaris and AIX, so we don't develop anything sold to individual users either. Nothing again Microsoft servers, but none of our customers use them.

Posted by: Scott Hanson on March 17, 2003 12:33 PM

a good article on open source and CIO decision making

CIO.com

Posted by: jjj on March 17, 2003 12:36 PM

Scott,

And I have nothing against MS, so we're even. ;-)

Most of the apps I've worked on (as a business analyst/project manager) have been using MS backends, and the sole excpetion was for Unix.

I've yet to see (or even hear of) Linux being deployed on the corporate desktop, though. I'm sure someone has but AFAIK most companies deploy on Windows.

Andrew,

It's true that web GUIs aren't great but then most corporate GUIs are pretty bad anyway. Few projects seem to pay enough attention to the GUI for it to make that much of a difference.

Normally GUI design seems to be handed over to a developer who just goes and whomps one up.

I'm not knocking good design, I'm just saying that I've never seen it elevated to the level of importance that would cause it to drive platform decisions.

I think what drives MS on the corporate desktop is the ubitquity of Office. Basically Office is such a standard that you gotta be able to run it on the desktop, and it's a lot easier to do that if you use Windows. Heck, Office is the killer app for me (and I know there are Linux equivalents, but I've put so much time into mastering those apps that the relearning curve would be a major impediment to me, plus I'm not aware of good Linux equivalents of Visio and Project).


Posted by: Kevin Brennan on March 17, 2003 01:07 PM

" Clearly it's time to mark my beliefs to market on this. Where did I go wrong? Where was the flaw in my chain of reasoning?"

Taking Paul David and Brian Arthur's nonsense seriously.

Posted by: Patrick R. Sullivan on March 17, 2003 02:42 PM

Most RAD tools solve problems I don't have any more. Just like the whole desktop windowing metaphor, they are dealing with issues of interfacing with something, not with getting to grips with that something itself. With RAD, that something used to be handled by mind numbing editing/coding/testing cycles, which made for bulk at the unit level. But my own problems don't come up at that Dick and Jane level, and I am perfectly capable of improving my unit coding/testing productivity with a good editor modularised away from the actual coding/testing. What about others with less experience than me? With RAD tools they will never get the "feel" for things, the road sense, that would lead them to higher levels of expertise. That's why it's valuable for a trainee to do menial work like carrying around boxes of tools.

Posted by: P.M.Lawrence on March 17, 2003 02:45 PM

There is a theory that, post WWII, Ford and Chrysler continued to exist because GM felt, for a complicated array of reasons, that it was in GM's best interest that they do so. (A not-so-complicated reason of course was that the New Dealers were still around).

Bill Gates may feel the same way. He gets innovation (and failure of such) onto somebody else's balance sheet, has an enemy to rally his troops around, & etc.

Posted by: a different chris on March 17, 2003 03:14 PM

Joel Spolsky says that it's only rational to write apps for Windows.
Neal Stephenson wrote about the Linux-friendliness of the cheap-hardware soup that Gates, Stallman, and Torvalds created. I don't know who's right.

URLs:
http://www.joelonsoftware.com/printerFriendly/articles/fog0000000296.html
http://www.cryptonomicon.com/beginning.html
http://www.joelonsoftware.com/articles/fog0000000051.html

Posted by: Vicki on March 17, 2003 03:40 PM

"There are good reasons for Linux to be winning the server market, but client-side?! Let's face it, you don't get good GUI in HTML, nor in Java."

Java applets make excellent GUI's for distributed apps. If you have any clue what you're doing, you can have a desktop quality user interface that runs on the vast majority of browsers, without dealing with the pain of JavaScript.

Posted by: Ken on March 17, 2003 03:59 PM

A few from thoughts from a bushel full:

Developers are in for the long term relationship. Possibly they don't trust Microsoft anymore and so they are seeking other avenues. Developers have to make choices in the face of huge uncertainity. Open Source offers an easier to understand relationship to developers. For a nervous developer the anti-trust trail exposed some very distressing behavior.

Microsoft is finding it extremely difficult to move it's OS hub from it's desk top center of mass onto all the places that CPUs are popping up - i.e. everything with electricity in it - without disrupting the cash flow from the existing business. The network is only one of these. They may well succeed, but in the meantime early to market opportunities abound for developers all over the landscape.

If it's true that developer mind share is shifting away from Microsoft that might or might not be a huge problem. Platform vendors don't innovate - they sit at the center of a hub of activity and capture those activities around them that threaten to aggregate sufficent momenteum to gain negotiating power. It's always been something of a blessing that Apple - which had to be more clever and could be lighter on it's feet for a number of reasons - would blaze a trail for Microsoft. The necessity of Office shipping on the Mac has always meant Microsoft could negotiate access to that IP.

Microsoft managed to convert the Mac from a competitor into a compliment quite a while ago. It's not clear that they can't do that with some other other threats as well. Killing off all your complements isn't a good strategy.

If nothing else Microsoft is persistent. It took for than a decade for them to catch up with 80% of the Mac's value proposition.

Posted by: Ben Hyde on March 17, 2003 05:44 PM

> Where was the flaw in my chain of reasoning?

There were two flaws, I believe. First:

> ever since Intel's chips began to eat the
> SPARCs for lunch

When Intel released 386, PC has finally become a Unix-compatible platform, which meant exactly one thing: BSD, which is by far the most stable OS to ever run on PC hardware. Granted, BSD is not a desktop OS, but it can work miracles on the server side. Which brings me right to your second flaw:

> What UNIX had going for it was that it ran on
> workstations

More importantly, Unix ran on SERVERS. So when 1990s saw huge growth in both "server farms" and workgroup servers, Unix (in its BSD and, later, Linux incarnations) was already there. Moreover, Samba (the software tool allowing integration of Unix machines into Windows networks) has been around since early 1994, so BSD and Linux servers could very easily be disguised as Windows servers. This had very interesting consequences -- network administrators desperate for better uptime actually began dumping Windows NT for BSD and Linux on workgroup servers without telling anyone. By the early 2000s, these administrators have grown up to be CTOs, and, voila, we have the whole Enterprise Linux movement, with IBM and HP leading the pack...

Posted by: Nikolai Chuvakhin on March 17, 2003 07:46 PM

"Where did I go wrong?"

Maybe you oversimplified a very large, very complex, and extremely competitive market, to put too much weight on economies of scale? [1]

Or you forgot Peter Drucker's dictum that every business with a dominant market position that is not protected by government plants the seeds of its own undoing?

Or as a person with market interventionist tendencies you fear a prospective monopoly a little more than is warranted by the facts, and you let your bias get the best of you?
(It brings to mind how a mere year and a half ago the Justice Dept protected consumers from the undue concentration of market power that would have resulted from the merger of United and US Airways, killing the merger. Now both are in bankruptcy.)

[1] After I graduated from law school, back in the last of the days when General Motors still had almost 50% market share, I had a job interview with GM's general counsel. He explained to me how the toughest part of his job was convincing the operating executives to keep GM's market share *under* 50%, to avoid anti-trust problems. He detailed it to me, since as I also had an interest in economics he figured I'd understand...

Because GM had such a big lead in market share, it was able to spread its fixed costs over many more sales. This gave it an inherent pricing advantage over the competition. In addition, GM had network advantages that no competitor could match, in the form of the only comprehensive product line in the industry that could fully take advantage of customer loyalty as consumers moved from needing one kind of car to another, plus supplier and dealer and service relationships that no competitor could match or build. In addition, the barriers to entry in the auto business were so high that new firms could not enter it.

Of course, the economies of scale and network advantages grew ever larger as the firm did, creating a competitive advantage that killed off competitors. This combined with the barriers to entry continually reduced the number of competitors -- from hundreds once-upon-a-time to then only three. And that only made the economies of scale and network advantages ever larger. And so on and on, in a cycle that really made it *hard* for the firm to keep market share under 50% by shifting enough profits into investments abroad.

And a half dozen years later it was losing billions and begging the government for protection from the Japanese.

The auto industry then was *much* simpler and slower than the computer industry is today, and the general counsel was not a stupid man who did not understand his company or the economics of anti-trust and monopoly. It's just that it is really, really easy to think one knows much more about a market than one does.

BTW, it may sound like I made all that up, but that conversation really and truly happened. And I didn't get the job. ;-(

Posted by: Jim Glass on March 17, 2003 10:22 PM

Miscellaneous thoughts from a software developer's perspective on why Unix has survived:

Most hard-core programmers are highly intelligent young men. Highly intelligent young men are also often idealistic, and dislike predatory monopolies like Microsoft for ideological reasons. Add to that the fact that many of them are students & don't have much money, and the Unix world offers them development tools that are as good as anything Microsoft has to offer, for free.

Unix originated as a hacker's platform and is much more developer-friendly than Windows. If you're one of the small minority of software developers who build desktop applications for sale, then it makes obvious sense to concentrate on the Windows market. Most people wouldn't freely choose to develop on Windows for any other resaon.

Most interesting new things in software development are happening on the web, where there is no overwhelming reason to choose Microsoft products - and security and reliability are fairly overwhelming reasons not to. The last time I checked on Netcraft (http://uptime.netcraft.com/up/today/top.avg.html), 49 of the 50 web servers with the longest uptimes were running Apache on Unix - and 48 of those were running variants of BSD. Also according to Netcraft (http://www.netcraft.com/survey/), Microsoft's share of publicly-accessible web servers peaked a year ago and has since dropped from around 35% to 28%.

Then we have internal corporate systems, the "dark matter" of the software development world - most commentators focus on the visible world of commercial desktop applications and the web, but internal corporate systems are where a large proportion of software developers, myself included, earn their living. Here there is some penetration of Windows servers for workgroup applications and some packaged stuff, but the industrial data-shovelling is done on mainframes or Unix. (As an aside - Brad thought Windows was going to make Unix and Macs go away. Ten years ago I thought Unix servers were going to make mainframes go away. Hasn't happened, isn't going to happen any time soon). Clients are increasingly web-based too - corporations hate the logistics and support costs of deploying desktop client applications. Desktops PCs are still normally Windows, but they are increasingly just used to run Office, IE and a few older client apps - for anything new there is a strong preference for building web front ends.

If we look at the small fraction of the software development world that is mass market desktop apps, then things look different. I see no evidence that Linux is anywhere near ready for mass desktop use in terms of quality and consistency of user interfaces - it's still a toy for techies. I like Java for back end development, but I've never seen a Java client app that wasn't slow and ugly. Which leaves Windows or the Mac as choices.

Windows - well, Microsoft's development tools are good but they lock you into Microsoft's proprietary non-standards. Microsoft appears to have overwhelming market share - but a very high proportion of that is corporate machines that just run Office and IE, whose users are never going to spend any money on independent software. And if, despite that, you actually do look like achieving any significant sales, Microsoft will ship an inferior clone of your product for free and kill your business.

The Mac - it's a niche, but it's Unix (which developers like) with a good user interface (unlike Linux). Apple's own development tools are good, plus you get access to all the marvellous free Unix stuff. Apple currently has a lot of buzz despite the hopelessly uncompetitive price-performance of its hardware - developer enthusiasm is responsible for a lot of that. And 3% of the entire worldwide market isn't that small a niche, particularly when you consider it's populated by enthusiasts who, unlike people with Windows PCs in corporate cubicles, have some ability and propensity to spend money on software. But if you actually look like achieving any significant sales, Apple will ship a not-so-inferior clone of your product for free and kill your market. Er ...

Posted by: Alan Little on March 18, 2003 02:59 AM

Two thoughts: One, I've noticed that the best of the ISV packages are written to be portable. They'll run on a lot of platforms, but they're only supported on one or two because the ISV or whoever actually markets the software doesn't want the added service costs. They support only numbers one and two in that particular market segment and there are a lot of segments-- from doorknob processors to always-up mainframes. If we could get beyond this limited service model we end-users would have many more operating systems to choose from and better price performance and lower cost of ownership.

Two, sales, especially resellers are still influential. They promote Unix and other non-Windows platforms because it's more profitable to them. People buy mostly what they're being sold, not what they've spec'd out.

Posted by: Dennis Slough on March 18, 2003 05:57 AM

Jim Glass: doesn't matter what happened later, the relevance is that many people believe that GM could have eliminated (just by buying them up if nothing else) Ford and Chrysler in the late 60's if it wanted to, and yet it didn't. Many people also believe that Microsoft also had the opportunity to eliminate Apple and passed- the most obvious support of that argument being that Microsoft actually propped up Apple in a memorable moment (the picture of Bill Gates' enormous image hanging, Ferengi-like, over Steve Jobs, and the ensuing boos will always be a classic Silicon Valley wacko moment).

But if you do move farther through history, you probably could say, like GM, the window for Microsoft (no pun intended) seems to have closed. What's hard to get through to the Heinlein-soaked libertarians is that in both cases it took somebody huge to make the difference: Japan, Inc in the automobile case and IBM's embrace of Linux in the case under examination.

Jousting elephants are never something one should view without a certain amount of concern. And yeah, that includes gummint anti-trust actions too.

Posted by: a different chris on March 18, 2003 06:55 AM

GM could not have bought up Ford and Chrysler, the Justice Dept would never have allowed it. I too was told of this fear of exceeding 50% of market share, in the late 60s, by a former GM employee who was a grad student at the U of Washington.

Somewhat to my amusement, since I'd had a job in high school detailing cars at a Datsun (now Nissan) dealership. Though not as amused as I later was to read this from Brian Arthur (aka, the Pied Piper of Lock-in), purportedly supporting his theories of path dependence:

"...in the early 1970's Japanese auto makers began to sell significant number of small cars in
the US. As Japan gained market volume without much opposition from Detroit, its
engineers and production workers gained experience, its costs fell and its products
improved. These factors, together with improved sales networks, allowed the Japanese
to increase its share of the US market; workers gained more experience, costs fell
further, and quality improved again. Before Detroit responded in a serious way this
positive-feedback loop had helped Japanese companies to make serious inroads into the
US market for small cars."

Which, according to Arthur and his colleague Paul David, should have been impossible.

Posted by: Patrick R. Sullivan on March 18, 2003 08:35 AM

I have been a happy user of Unix for some years now and I have experience with Linux, FreeBSD and Solaris. The switch from Windows to Linux at the time was somewhat of a revelation. I directly had the feeling that the architecture behind Unix was superior to Windows (file system, memory management etc) and I also had the feeling that you could do more with Unix than with Windows, with lots of free tools available. Although MS has advanced with its WindowsXP offering in terms of usability and stablilty, it still can't beat a Unix system like Linux or FreeBSD. My opinion is that MS-Windows is bound to lose in the long run (5-10 yrs) as even the hold of MS on the desktop will slowly but surely weaken, due to the combined power of Unix and open source software (like OpenOffice).

Posted by: Nescio on March 18, 2003 01:43 PM

"What's hard to get through to the Heinlein-soaked libertarians is that in both cases it took somebody huge to make the difference: Japan, Inc in the automobile case and IBM's embrace of Linux in the case under examination."

That's a myth, the truth was the reverse. "Japan Inc.", i.e. the famous MITI, which set Japanese industrial policy, decided the Japanese car makers were *incapable* of competing in the USA and tried to legally *ban* them from doing so until it could form an exporting cartel.

This infuriated Soichiro Honda, who was trying to build his cycle and scooter company into a car maker and had decided that he needed access to the American market to do it. So he effectively welded two golf carts together to meet the definition of a car under the law, and shipped it the US before the "grandfather clause" export deadline in it, which nobody was supposed to meet. That broke the cartel.

Honda then started exporting to and then manufacturing in the US. It was the smallest Japanese car company, a start-up in the auto business. After just a few years it had the #1 selling car in the US, savaging GM more than any other US company. The other Japanese car makers followed its lead.

So reality is the reverse of the myth. "Japan Inc." totally misjudged the situation and tried to block Japanese car exports to the US. An entrepreneur defied Japan Inc., broke its cartel and took on all the US car companies as well, as a virtual start-up. He had great success and created the Japanese auto industry in the US as we know it -- over the objection of the Japanese gov't planners. The biggest of the US companies, GM, was pummeled the most -- by start-up level entreperneurial competition. And Soichiro Honda despised MITI and Japanese industrial policy until the day he died.

We might also remember that the great IBM, which kept its market power through one of the biggest anti-trust suits of all time (certainly the longest) and which was anointed by many as the US's answer to the mighty "Japan Inc." around 1990, when it was the most profitable US corporation of all time, was promptly thereafter devastated by a bunch of hardware start-ups that emerged literally out of garages and tents, and a little software company whose programmers slept on the floor.

Posted by: Jim Glass on March 18, 2003 08:41 PM

I can't speak for the rest of the world, but, for me, Japan Inc. was not the MITI, but the whole of Japan. The MITI was at most part of it, and not the part that understood business at that.

DSW

Posted by: Antoni Jaume on March 19, 2003 11:38 AM

Jim, these entrepreneur's also unmasked the large amount of consumer welfare being lost. It sure is good to be king of the hill when economies of scale are at play... So whose money is it?

Posted by: Stan on March 19, 2003 01:03 PM

Another reason why Apple is still in business. Commitment to good corporate oversight:

Apple Computer Isaid that former U.S. Vice President Al Gore had joined the company's board. "Al brings an incredible wealth of knowledge and wisdom to Apple from having helped run the largest organization in the world -- the United States government," Steve Jobs said. "Al is also an avid Mac user and does his own video editing in Final Cut Pro."

Gore had been particularly impressed with Apple's Mac OS X operating system and the company's commitment to the open-source software movement.

Posted by: bakho on March 19, 2003 04:06 PM

"It sure is good to be king of the hill when economies of scale are at play"

OTOH, it's pretty damn bad to be locked into "economies of scale" when a smaller innovator changes the processes in your market, so your "economies" of scale become deadweight costs of scale that you can't escape.

IBM went from making the largest profits in corporate history to making the largest losses in corporate history in three short years. It hadn't lost any scale in the meantime.

Posted by: Jim Glass on March 19, 2003 08:27 PM
Post a comment