Softpanorama (slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-)

Nikolai Bezroukov. Portraits of Open Source Pioneers

For readers with high sensitivity to grammar errors access to this page is explicitly prohibited :-)

Chapter 4. Part 3:
Burst of Internet Technobubble

• Who was behind Linux gold rush
   
• Cuckoo effect: Linux as a major Microsoft ally in fighting Unix vendors

  • Linux helps Microsoft to win federal lawsuit; the software marketplace "lock-in-effect"

  • Linus Torvalds as an example of the closed-source style development of open source software

  • Attacking proprietary Unix market

  • Linux kills SCO
     

In June 2004 ItWorld.com conducted an informal poll. It consisted with just one question: "Has open source sold its soul?". Here is the text:

Do you think that the embrace of open source by big companies has killed
the movement's idealistic drive?

* No - the movement is still going strong, and is transforming corporate
cultures from the inside
* Yes - the revolution has turned into just another way to make money
* A lot of it's about cash now, but you can still set up your own
project that reflects your ideals
* Hey, I've always been in it for the money

Let us know!
http://itw.itworld.com/GoNow/a14724a105841a75978418a14

Here are the results as of July 1, 2004:

POLL: Do you think that the embrace of open source by big companies has killed the movement's idealistic drive?
49.1%  No - the movement is still going strong, and is transforming corporate cultures from the inside
25.9%  Yes - the revolution has turned into just another way to make money
17.2%  A lot of it's about cash now, but you can still set up your own project that reflects your ideals
7.8%    Hey, I've always been in it for the money

While definitely unscientific, the results above suggest that more then a half correspondent think that open source (to be more exact commercialized GPL-based software including Linux, but excluding Apache and all BSD-based projects like FreeBSD, OpenBSD, NetBSD, etc) became a puppet in big corporate games ("sold its soul"). This part of the chapter contains my investigation into the process. 

Who was Behind Linux Gold Rush

In the cover story "Net Money Game: How top financial firms reaped billions, while investors got burned" Business Week noted that for years, venture capitalists declared that their mission was to build sustainable businesses. But when Linux IPO mania hit, they began rushing companies onto the public market with just promising business plans, but not sustainable business model. 

Unfortunately Linus Torvalds and Linux companies became part of the "make money fast" scam played on unsuspecting public by investment banks and venture capitalists firms; Linus mainly like a pawn in well trained hands of PR staff of  investment banks, companies like Red Hat as a willing conspirators.  My experience leads me to believe that easy money from VC are big stimulus for con-games to be played on investors, and it really doesn't take that much money to make a fake Web site extolling virtues of non-existent or weak product on the Internet.  Like PR people like to say its all about perception, not about reality.

Maybe all of the baby-boomers who bet (and lost) their retirements on the Linux startups IPOs will be happy to know that such mania are periodic event in human history since the tulipobubble. 

Long before anyone ever heard of  Linux and Linux startups like Red Hat, VA Linux, and Caldera that have soared during the technobuble, there were tulips. In 1593, a Vienna botany professor brought a collection of unusual bulbs to Holland that had originated in Turkey. That triggered events that led to one of the most spectacular get-rich-quick binges in history.  Over the next decade, the tulip became a popular, but expensive item in Dutch gardens. These flowers were stricken with a non-fatal virus known as mosaic. The virus caused the tulip petals to develop colored stripes which led to wild speculation in tulip bulbs. Popular taste dictated that the more bazaar the bulb, the greater the cost. Bulb prices rose out of control. The more expensive the bulbs became, the more people viewed them as smart investments. The ordinary industry of the country was dropped in favor of speculation in tulip bulbs.

Everyone imagined that the demand for tulips would last forever and that people all over the world would pay any price for them. People who claimed prices could not possibly go higher watched their friends make enormous profits. The temptation to join them was irresistible and few Dutchmen sat on the sidelines. In the last years of the tulip craze, which occurred from 1634–1637, people bartered their personal belongings, land, jewels and furniture to obtain the investment that would quickly make them rich. As happens in all speculative crazes, prices eventually soared so high that some people decided they should sell.

Soon others followed with a snowball effect.  The government tried to calm panic stating officially that there was no reason for tulip bulbs to fall in price – but no one listened. Dealers went bankrupt and most bulbs became worthless – selling for no more than the price of an onion.

Here is how Business Week describes tulipomania:

To citizens of 17th century Holland, tulips became the most desirable item to have. Around 1624, the Amsterdam man who owned the only dozen specimens was offered 3,000 guilders for one bulb. While there's no accurate way to render that in today's dollars, the sum was approximately equal to the annual income of a wealthy merchant. (Rembrandt received about half that amount for painting The Night Watch a few years later). Bulb prices rose steadily throughout the 1630s, as ever more speculators wedged into the market. Farmers mortgaged whatever they could to raise cash to begin trading. In 1633, a farmhouse in Hoorn changed hands for three rare bulbs. By 1636 any tulip--even bulbs recently considered garbage--could be sold off, often for hundreds of guilders. A futures market for bulbs existed, and tulip traders could be found conducting their business in hundreds of Dutch taverns. Tulip mania reached its peak during the winter of 1636-37, when some bulbs were changing hands ten times in a day. The zenith came early that winter, at an auction to benefit seven orphans whose only asset was 70 fine tulips left by their father. One, a rare Violetten Admirael van Enkhuizen bulb that was about to split in two, sold for 5,200 guilders, the all-time record. All told, the flowers brought in nearly 53,000 guilders. Soon after, the tulip market crashed utterly, spectacularly. It began in Haarlem, at a routine bulb auction when, for the first time, the greater fool refused to show up and pay. Within days, the panic had spread across the country. Despite the efforts of traders to prop up demand, the market for tulips evaporated. Flowers that had commanded 5,000 guilders a few weeks before now fetched one-hundredth that amount.

It was a similar epidemics of greed that sent in the end of XX century the classic rule of thumb for investment banks "Don't take a company public until it has three quarters of profits"  to the waist basket. You can spoke about lower ethical standards, but in reality it was classic example of greed that broke the system. Actually Linux IPO gold rush was just a part of a bigger gold rush, the Internet bubble.

According to Thomson Financial,  taking public hundreds of companies that have no sustainable business model and no profits, investment bankers took in $2.1 billion just in underwriting fees since 1997.

It is interesting to note that much more then $2.5 trillion later has disappeared from the market since the February 2000 in a process that was later called "Dot com bubble bust". That means that never before investment companies managed to help so many investors to lose so much money so fast. 

For example, in May 2001 of the 367 Internet firms taken public since 1997 that are still trading, the stocks of 316 were below their offering prices. Along with classic example of VA Linux one can find hundred, if not thousand of other no less striking examples. Some even manage to beat VA Linux in the speed of downward spiral: Pets.com Inc. managed to go from its $66 million IPO to closing its doors in just 10 months.   

At the same time several America's leading investment banks reaped millions in fees. And Linux IPOs played a minor but still historically important part in this scheme. Only in retrospect it became clear that they were playing a different, more sophisticated and cruel money game than the investors. Investors that naively expected on stocks rising from their IPO levels to make money were burned, but investment bankers took their cash up front and if the stock went down, they already had their money in the bank.

For example , for poor Pet.com mentioned above took Merrill Lynch  took in fees of about $4.6 million: that is almost 5% of approximately $100 million Merrill got since 1997 for  Internet IPOs. 

Later when class action lawsuit was brought against both investment banks and tech companies that were brought to public using laddering (scheme in which larger pools of shares are allocated to investors who promise to take bigger stakes after the stock hits the open market. In addition, those who got large number of IPOs shares were compelled to give extra compensation to the banks, sometimes through inflated commissions on other trades. Later, after the so-called "quiet period" that follows IPOs, analysts who worked for the banks issued favorable research to fluff up the stocks, whether they were worthy or not).

U.S. District Judge Shira A. Scheindlin decided the cases could go forward, rejecting a bid by the investment banks to dismiss the case. Scheindlin reviewed more than 1,000 suits and found that investors had presented "a coherent scheme" in which the banks joined with Internet companies to defraud the public by hiding secret deals and analyst conflicts to artificially inflate new shares and deliver payoffs to insiders. If the allegations are true, she wrote in her 238-page ruling, "this scheme offends the very purpose of the securities laws." See InformationWeek IPO Fraud Tech Company Insurers To Pay At Least $1 Billion In IPO Fraud Case June 27, 2003 for more details:

NEW YORK (AP) -- Hundreds of companies that staged hot initial public offerings during the tech boom will pay $1 billion to investors under a tentative partial settlement announced Thursday, and cooperate in ongoing litigation against 55 brokerage firms accused of funneling huge payoffs to insiders through secret deals.

The massive case involves 309 separate suits filed against 55 investment banks, more than 300 companies that went public between 1998 and 2000, and an unspecified number of their individual corporate officers and directors. The total number of defendants could be more than 1,000; companies involved include Global Crossing, MP3.com, Ask Jeeves Inc., and Red Hat Inc.

The proposed settlement guarantees the plaintiffs will receive at least $1 billion, to be paid by the tech companies' insurers, said Melvyn I. Weiss, chairman of a committee of attorneys representing the investors. The plaintiffs hope that settling with the companies that issued the stock--who they say knew or should have known about the alleged misconduct--will strengthen their position as they continue negotiating with the investment banks.

"We have always been of the mind that the primary target in this case is the underwriting community," Weiss said. "This gives us a huge booster shot in our case against them."

The plaintiffs are looking to recover "many billions" of dollars from the investment banks, Weiss said. The firms, including J.P. Morgan, Credit Suisse First Boston, Morgan Stanley and Smith Barney, are accused of plotting to artificially inflate the value of IPO stocks through a practice called "laddering," in which larger shares are allocated to investors who promise to take bigger stakes after the stock hits the open market.

In addition, some customers who invested in IPOs were compelled to give extra compensation to the banks, sometimes through inflated commissions on other trades. Later, after the so-called "quiet period" that follows IPOs, analysts who worked for the banks issued favorable research to fluff up the stocks, whether they were worthy or not.

As part of the settlement, the tech companies have agreed that any claims they might have against the investment banks will be assigned to the plaintiffs' class. Their cooperation may also help speed the discovery process, Weiss said.

"They participated in road shows, they had conversations with the underwriters," Weiss said. "They may have been misled as to what compensation the underwriters were receiving. ... all of this is important to us."

Most of those involved have approved the settlement, Weiss said, which also must be approved by the court.

No funds will be paid to investors until the case against the banks is resolved, however. If more than $1 billion is recovered from the banks, the tech companies' will not have to pay anything, the lawyers said. If the award is more than $5 billion, the tech companies and their insurers will be able to recover expenses.

The settlement was reached after more than a year and a half of negotiation between lawyers for the investors, the tech companies and at least 42 primary insurers. Attorneys involved agree it is among the most complicated cases in U.S. history.

"This is by far the most complex securities litigation that has ever been brought, and the settlement process is equally complex," said Jack Auspitz, a lawyer with Morrison & Forester, who represents 30 to 40 of the Internet companies.

In February, U.S. District Judge Shira A. Scheindlin decided the cases could go forward, rejecting a bid by the investment banks to dismiss the case.

Scheindlin reviewed more than 1,000 suits and found that investors had presented "a coherent scheme" in which the banks joined with Internet companies to defraud the public by hiding secret deals and analyst conflicts to artificially inflate new shares and deliver payoffs to insiders. If the allegations are true, she wrote in her 238-page ruling, "this scheme offends the very purpose of the securities laws."

The Securities and Exchange Commission and the National Association of Securities Dealers conducted an extensive investigation of Wall Street's dealings in IPOs. Regulators have recommended strict limits on several questionable practices popular during the tech boom, including laddering.

Ten of the largest investment banks agreed to change their research and IPO practices in a $1.4 billion settlement reached with regulators earlier this year.

But it's too simplistic to blame investment banks alone. This wave of greed  affected other players too. Venture capitalists had a similar spot in the food chain. In exchange for footing the early costs of a startup, they got stock for pennies a share. That means that even if shares dove to a dollar per share after the offering, they can make a huge profit. And they probably played the second role after the investment banks in creating Linux gold rush and definitly deserve a share of blame for the Linux bust.

Institutional and individual investors are guilty  too, because many got caught up in the Linux gold rush.  After all the risks detailed in their offering documents often included phrases such as "unproven business model" and "may not continue as a going concern." 

That means that from now on most investors may shy away from investing in Linux startups, good and bad. Already, in early 2001 the number of initial public offerings for tech companies has slowed considerably. And this money drought for Linux startups may last several years. 

As I mentioned before the key idea was laddering: inflating the stock price at the beginning whatever it costs, allowing major players to cash out profits and then leaving the company to struggle with consequences. It was like building a bank branch in order to rob it later. That also means that high level executives in those companies were willing participants in a rather dirty game. In a free-market economy, companies are supposed to fail because of the recession of wrong business decisions or both. But here we essentially have a preprogrammed failures after the short initial inflated growth, hype and other dirty tricks due to sleazy practices, if not outright fraud. Off-balance sheeting and other "creative accounting" tricks became the requisite way to boost earnings.  Slowly but surely financial truth disappeared and reported earning became an exercise in science-fiction.

Many naive investors, especially in open source community, embraced those stock because they believe that they contributed to a diversity on the market the suffered from Microsoft domination. They wanted to participate fairly in the creation of an alternative to Microsoft, but instead they became victims of unscrupulous "make money fast" adventurists and later saw their investments disappear. 

At the end Linux companies IPOs did not pass a smell test. Even if company officers and investment firms and their lawyers can ultimately show that they following the letter of the law, after collapses of companies like VA Linux too many people remember those IPOs as yet another elaborate financial hoax. 

Privately, some investment bankers admit the wrongdoing. But with the amount of money they got don't bet on the lessons leaned. In the cover story Net Money Game: How top financial firms reaped billions, while investors got burned" Business Week provided the following data:  

...First, the successes. Morgan Stanley Dean Witter (MWD ) led the IPOs of many Net companies that have tumbled sharply from their offering prices. It reaped total underwriting fees of $350 million since 1997, according to Thomson Financial. But the firm has a respectable average return of 20% because of big hits with Net security systems developer VeriSign Inc. (VRSN ) and storage solutions provider Brocade Communications Systems Inc. (BRCD ) On the venture-capital side, Breyer's Accel Partners backed 30 companies that have gone public since 1997, and those upstarts have averaged returns of 55% for public shareholders, according to Venture Economics. They include audio and video software provider RealNetworks (RNWK ) and networking equipment maker Redback Networks Inc. (RBAK ) VC firms Sequoia Capital and Kleiner Perkins Caufield & Byers also have strong track records, with average returns of 54% and 23% respectively.

Few on Wall Street, however, have strong returns. While taking public hundreds of Net companies that have proved to be losers, investment bankers took in $2.1 billion in underwriting fees since 1997, according to Thomson Financial. The biggest names in the business led the charge. Goldman Sachs, arguably the most respected firm on Wall Street, reaped fees of $360 million by leading or co-leading the IPOs of 47 Net companies since 1997. Now, 38 of those companies have stocks trading below their offering prices, and two more have shut down. The average return for investors is -16%. Relatively, that's not bad. Credit Suisse First Boston, Robertson Stephens, Bear Stearns (BSC ), Lehman (LEH ), and Merrill Lynch all averaged returns of -45% or worse, according to Thomson Financial's figures. "Clearly, lots of very high-risk companies went public, and the investment bankers didn't show a whole lot of discipline in turning down deals," says E. David Coolidge III, chief executive officer at William Blair & Co., a 65-year-old Chicago investment bank that largely sat out the Net boom.

Some bankers take issue with the numbers. Execs at CS First Boston contend that their track record should be measured only since star tech banker Frank Quattrone and his team moved to the firm in mid-1998. If you measure the years 1999 and 2000 and include the IPOs that CS First Boston considers Net companies, the firm's average return is -55%, which is not as low as the returns they claim that Morgan Stanley and Goldman had. "The performance of Internet stocks has been pretty ugly for all of us," says John Hodge, managing director for U.S. technology corporate finance at CS First Boston. "But we did more deals and our average returns are better."

Lack of insight. If the results have been ugly, investment bankers point out that they were struggling with some of the same issues as investors. Companies were going public so early that Wall Streeters had a hard time differentiating winners and losers. "Because the market was clamoring for these companies, we had to make judgments earlier in companies' life cycles," says Goldman's Koenig. "The market was interested in investing in these companies at a stage when proof of viability was impossible to get."

Ranking the Investment Bankers How have investment banks performed in taking Internet companies public? Here's a look at the average returns for the Net companies taken public since 1997 by each bank.
Morgan Stanley & Co.	20%
Salomon Smith Barney	-14%
Goldman Sachs & Co.	-16%
Deutsche Bank AG	-41%
CS First Boston	-41%
Lehman Brothers	-59%
JP Morgan	-59%
Bear Stearns & Co Inc	-65%
Robertson Stephens	-65%
Merrill Lynch & Co.	-82%
Data: Thomson Financial Securities Data, Business Week

The track record suggests some investment banks didn't have any more insight into the prospects for certain companies than the average investor. Lehman Brothers Inc. led the IPO of U.S. Interactive Inc., a Net consulting firm based in King of Prussia, Pa., in August, 1999, at $10. The stock surged to $76.50 in January, 2000, and then began a steady slide. Lehman analyst Karl Keirstead issued buy recommendations seven times between February and early September, 2000, as the stock dropped 92%, to $6.50. Only after the company disclosed on Sept. 20 that it would miss third-quarter financial estimates did Keirstead downgrade the stock--and then merely to "outperform." Since then, U.S. Interactive has dropped to 50 cents a share and, in January, filed for bankruptcy. Sharon Wilson, an accountant in Oregon who's nearing retirement, was one investor who got nailed. She began buying U.S. Interactive at $73.15 a share and ended up paying about $13,000 for 5,000 shares. Now, Wilson is out at least $10,000. "Many of us are ready to shoot most of the analysts out there," she says.

Lehman and Keirstead didn't fare too badly, though. The firm received fees of about $2 million for managing the books for U.S. Interactive's IPO, according to Thomson Financial. Even without U.S. Interactive, its track record isn't one to brag about: For the 15 Net companies it took public that still trade, investors have lost an average of 59% from the IPO price. They include online community developer Talk City (TCTY ) (down 99%), health site HealthCentral.com (down 97%), and ad tech firm Mediaplex (down 95%). Keirstead did not return telephone calls seeking comment.

Other Wall Street heavyweights played the same game. In May, 1999, Salomon Smith Barney led the IPO of Juno Online Services Inc. (JWEB ), a New York company that provided free e-mail service and later started offering free Internet access. In June, Lanny Baker, Salomon's Net analyst, issued a report with a buy recommendation on Juno, and by December, 1999, the stock hit a peak of $87, up from its $13 offering price.

But skeptics saw a classic Net-craze company. "There was no there there," says Steve Worthington, a portfolio manager at Barbary Coast Capital Management who sold Juno's stock short. "Giving away stuff for free never seemed like much of a business." Over the first half of 2000, investors came to agree, as Juno shares slid to less than $10. Even after that drop, Baker issued another report on June 1 that recommended investors buy the stock. While the stock recovered briefly to close at $12.44 on June 7, it then went into a slow, steady slide. On Aug. 2, with the stock again below $10, Baker downgraded the stock to neutral, and it now trades at about $1. Salomon defends its support of Juno. "We believed in Juno," says Christenson. "We were wrong."

... ... ... 

The truth is that some investment firms do not need any insight and do not care about the future of the company they brought to market. They want their money "now" and get them. Essentially this was the latest reincarnation of the famous "After me deluge" principle, if you wish. As Washington Post noticed in the paper Collapse of Dot-Coms Stifles Tech Innovators:

Caught in the middle in these uneasy economic times is the inventor, the hero of the new economy.

Some technologists complain that they have had trouble raising money for ideas that a year ago would have been shoo-ins for funding. That isn't to say they are out of work. Jobs for skilled tech workers are still plentiful. Pay is still good. But more than a few engineers are now finding themselves doing what amounts to virtual construction work, such as creating bill-payment or shopping systems, or securing corporate networks.

"Last year 80 percent of the technical jobs were creative, new. Now it's, like, 80 percent practical," said Paul Villella, president of HireStrategy.com, a Washington area online recruiting firm. "We've seen a complete reversal in the job market. It's not fun, innovative stuff anymore. It's more 'we-gotta-get-it-done and make-it-work' type work."

That transformation hasn't been good for morale.

Linux Cuckoo-effect
Linux as a major Microsoft ally in fight Unix vendors

It's well known that Microsoft was an implicit force behind the Linux meteoric rise. Without deep need for the alternative OS on PC Linux might never leave the world of amateur OS designers. But what is the most interesting observation is that Linux  (in free/open software development world) was pretty similar to Microsoft Windows in the world of commercial software. And Linux in turn helped Microsoft in several non-obvious ways. One of the was fighting off federal lawsuit.

Linux helps Microsoft to win federal lawsuit;
the software marketplace implicit "lock-in-effect"

This was a long and complex lawsuit and it's better to find all detail of it elsewhere. We will discuss it only as for the paradoxical role of Linux in this lawsuit. In short, Microsoft was accused in destroying Netscape and abusing its monopoly on the desktop. In reality all monopolists tend to abuse their power and Microsoft behaved more like a benevolent dictator then a destroyer of the marketplace, at least it behaved much better then IBM two decades ago (in the mainframe era). This is especially true regarding Oracle, one of the companies who (many suspect) where one of the major driving forces behind this federal lawsuit.

Actually competition to a certain extent survived and both Oracle and Linux were pretty convincing illustrations of this fact. The success of Linux also undermined the case for Netscape. It's clear that the real factors behind Netscape demise were more complex than Microsoft bundling it's IE with Windows. I strongly doubt that Netscape would survive even if Microsoft implemented a special selection screen (Netscape or IE) during the Windows installation.

All-in-all it's pretty stupid to portray Microsoft users as poor, misguided lemmings — people who haven't had their eyes opened to the joys of Netscape, Linux or the Mac OS, or who are forced by Microsoft to use inferior Office suit, browser, media player and/or messaging software because they come with Windows. But who is forcing whom? Who forced me to use Microsoft FrontPage for creating those pages despite this site open source orientation ? Nothing prevents me from installing Dreamweaver or whatever better HTML editor I can find on the market. I just did not find an HTML editor that can convince me to switch from FrontPage despite the fact that Dreamweaver is the standard HTML editor for  my "day-time" employer. And nothing prevents me from uninstalling Windows at home and using only Linux or Solaris or OS X as a client OS.  Moreover I do use Linux as a home server (despite the fact that I have an unopened box of NT4 server in the closet a license for Windows 2003 server), so that would even make the environment more homogeneous and much easier to manage. The problem is that I do not see any evidence that benefits would outweigh shortcomings of such a solution.

Actually with Office XP it looks for me that Microsoft is behaving not like a monopolist, but like a company that is struggling to keep pace with a competing software in the saturated marketplace and is pressed by competition to slash prices (you can get OEM Office XP standard for less than $200. That means that Word, Excel and PowerPoint and Outlook are less than $50 each -- which is a typical shareware price; I doubt that you can find alternative shareware products of comparable quality and functionality for such a price).  And if somebody want to use Word Perfect instead of Word nothing in Windows prevents him to do that even without being a professional programmer, like many lawyers know pretty well ;-). Actually Dell installs it on cheaper PCs instead of MS Office to cut costs. 

Linux role in this lawsuit was proving (by government own standards) that there is a competition in the OS market and that Microsoft does not artificially forcing  "lock-in" on consumers -- if somebody wants to use alternative OS or alternative applications suit like many Linux users prefer, he or she definitely can.  In a recent paper David S. Evans and Richard Schmalensee suggested that classic antitrust law might not be the best tool to deal with these issues. The two economists, who have written papers and legal filings supporting Microsoft's position in its antitrust case, said that government should act to keep companies from abusing their monopoly power to undercut competitors, but that "the application of antitrust principles should take account of the important ways new-economy industries differ from traditional ones. "In particular, they wrote, if a dominant company can prove that there is healthy competition despite its commanding share of a market that has winner-take-all characteristics, that fact "should be a defense to claims of predation."

The similarity of Microsoft rise to power and Linux rise and subsequent Linux IPO gold rush were pretty interesting phenomena if you take into account that Linux was not the best offering in the open OS arena: it seems that like in case of MS DOS some other non-technical forces helped Linux to achieve its prominence in the free/open source space. Here Linux could be even more valuable to Microsoft as a legal defense, but Microsoft lawyers did not use this argument ;-). And whatever the answer in the Microsoft case, there is a growing realization that Microsoft may be only the most visible symptom of a bigger and more severe problem afflicting software and Internet development.

It seems there are some implicit economic and social forces/factors that help to create dominant companies in the software space be it a commercial space (Microsoft) or open source space (Linux). And it's easy to see that the same forces that raise Microsoft to prominence are active in open/free software microcosm and helped to get Linux to the same position (unfairly as many BSD lovers can attest).  The only problem is that Linux users are a minority and probably will remain a minority in the decade to come.  So the real question is to what extent they deserve "an affirmative action" campaign from the government as well as from major UNIX vendors like IBM,  Sun and HP.  The whole complexity of the "affirmative action' issue toward minorities is now bound to be played in tech. arena.

This implicit "customer-lock-in effect" similar to Microsoft vs. Netscape we saw in the fight between FreeBSD and Linux.  Actually for most users Linux is synonymous with Red Hat, the fact that illustrates this "lock-in" effect even more convincingly.  It creates a different kind of barrier to entry for newcomers — not the cost to the company of entering the market, but the cost to create an infrastructure. You may think that bloated Linux kernel is inferior to the simpler and more elegant BSD kernel, but just count the number of Linux books vs. the number of FreeBSD books and the number of native Linux applications vs. native FreeBSD applications. Than you might understand better the potential cost for consumers willing to make a switch to technically superior FreeBSD camp. Until Apple entrance into the game the application space for Linux was much more developed than FreeBSD space and that creates serious difficulties for the potential OS switchers. Look how many specialists that previously almost exclusively use FreeBSD are now flirting with Linux :-)

There can be even factors unique to the world of software that makes it a field were "natural monopolies" can flourish. Some economists suggested that there is a special "software networking effect": the value of a software  grows geometrically as more and more people use it.  This "software networking effect" can be seen in technologies like Napster and America Online's Instant Messenger. In the latter case, once people realized that they can communicate online in real time using this protocol, it became a must-have, and its use exploded, rapidly bringing AOL a near-lock of a market that Microsoft now is struggling to break into.

Put all of these together and you get an environment in which benefits the biggest and the most ruthless player and where consumers themselves reward the dominant player to increase its lock on the market irrespective to the technical quality ("good enough" effect).  The result is a number of markets dominated by the biggest player.  And this effect is not limited to software. Intel grabbed almost 80% of the PC CPU market due to the same socio-economic forces. And recent announcement by Compaq Corp. that it would stop offering customers its own Alpha processors by 2004 means that even a better product supported by a powerful player does not have much chances on the marketplace. And while people often talk about Microsoft monopoly, a very similar monopoly exists for Oracle, which dominates the database market, Sun, that dominated Unix server market and America Online, the biggest ISP.

Linus Torvalds as an example of the closed-source style development of open source software

Talented managers and managed development is the key to a quality large scale software development. No anarchistic gang of developers can really compete with the real organization headed by a talented manager. And that's the lesson of development of Linux. And I think that paradoxically one of the achievements of Linus Torvalds is that he from the beginning tend to use classic management style of closed-source project in open source environment.

And his brave and smart idea to promote commercial distribution of Linux and commercial distributors instead of Gosplan (of USSR fame) style Free Software Foundation speaks volumes in this respect.   That was essentially a turning point of the development: open source projects managed by commercial companies with the explisit aim of chepaning or even outsourcing thier operations.

This idea was probably best formulated by Fred Baker -- a "fellow" at Cisco and the former head of the Internet Engineering Task Force. As reported in Open Sourcers Shy From Criticism he noted:

And he did have a point -- many open projects are managed by commercial companies, and it shouldn't be inflammatory to say so. It is also to be expected that a certain proportion of any popular movement will be zealots, and as bigoted as such folk tend to be.

Attacking proprietary Unix market:
 Linux destroys SCO and damages Solaris on Intel

Paradoxically Linux and Microsoft's can be considered close allies that both try to dislodge classic Unix as the dominant OS within the next decade. The shift had started with smaller businesses moving to install Linux instead of Unix workstations and for file and print services, replacing lower-end Unixes such as Solaris, AIX and HP-UX. Both Linux and Microsoft gain ground as IT managers try to take advantage of the cheaper maintenance of Linux and Windows in comparison with proprietary Unixes.  They also want to replace more expensive Risk servers with cheaper Intel-based servers.

As major software vendors declare support for Linux it becomes a real threat for Risk boxes on low end and partially on midrange. The other reason is that Linux is close enough to other Unixes and transition is more or less smooth and does not require extensive retraining of staff.

In 1999 HP declared that it is was no longer committed to it PA RISC architecture, moving instead to Itanium, making HP the first vendor that more or less openly embraced Linux as an alternative to its own proprietary version of Unix (HP-UX).  They even hired a Linux evangelist Bruce Perens to help them to destroy HP-UX market share ;-). He later was fired, but still managed to inflict a lot of damage. 

Linux kills SCO Unix

Just before going down in flames, Caldera that was traditionally considered to be a second after Red Hat Linux distributor, managed to acquire SCO and essentially killed SCO Unix. Although not that successful during the IPO as Red Hat, Caldera was a pretty rich company even before IPO due to Microsoft settlement money ($400 million in cash) and actually did not need IPO to raise money as badly as Red Hat and VA Linux.  Still greed is greed...

Paradoxically in some sense those extorted due to successful DR Dos lawsuit money from Microsoft turned to be a rather shrewd Microsoft investment. They continued working for Microsoft even after they moved into Caldera bank account: they helped to kill the Server Unix and Professional Services division of Santa Cruz Operations (SCO), the previous leader of commercial Unix on Intel. Even when Linux fashion wave started, SCO was the leading Unix distributor holding 37% of the Unix market with 313,000 software license shipments in 1999, compared to Sun's 22% of the market. But in 2000 Linux managed significantly undermined SCO market share and especially the mind share. That's why the company fall victim of this new competitor.

That event showed one again that contrary to popular belief Linux is not cannibalizing Microsoft share on desktop and has difficulties fighting with Microsoft on Intel servers due to Microsoft dominance on the desktop.  Actually the share of Linux desktops may well be flat (less than 5%) or even shrinking percentage wise as most new PC are still shipping with Windows preinstalled. As a desktop Linux is just sitting somewhere on the periphery serving special needs of certain small user subgroups.

At the same time Linux proved to be a very potent threat to it's Unix proprietary friends as a low level to midrange departmental server and WEB server and in those roles it has been replacing products such as SCO Unix and Solaris on Intel (as well as proprietary architectures by IBM, HP-UX and Sun). In the third quarter of 2000 SCO reported  loss of $19.24 millions, three times that expected, and was hemorrhaging cash at a speed that cannot be sustained.  That make the company ripe for an acquisition and in August, 2000 after two years of  free fall of stock price the company was acquired by Caldera. Actually Caldera  acquired not all company but just SCO's server software division and professional services division. The other part of the company remained independent, got some money and said it will use the new funds to invest in its Tarantella software, which it retains along with its Unix intellectual property rights. SCO surviving part was very tiny: Tarantella revenues amounted to just $2.5m in the third quarter of 2000.

SCO's server software and professional services divisions has been bought for 17.54 million shares (28 per cent of the company) plus $7m cash. At this time Caldera's share price was around $7, which made the deal worth $130 million. Stock in both companies has been in free fall for a long time. SCO's stock price has fallen from over $31 to under $4 since the start of the 2000. Caldera shares have fallen from a top price of $33 in April to $7.

Caldera decided to continue to run the OpenServer Unix-on-Intel operating system, which generated just $11.1m revenue in the third quarter of fiscal 2000, as a separate unit through new holding company Caldera Inc, but soon changed its mind and reorganized under the name of SCO.  The acquisition opened an interesting possibility for Caldera in view of previous SCO involvement with the Monterey project, the possibility which was not entirely dissimilar to the Caldera move to acquire DrDos and sue Microsoft. The potential problem was that SCO crown jewels:  proprietary clustering technologies for political reasons were pushed IBM to Linux community.  This threat later materialized, but the story of SCO lawsuit against IBM is beyond the scope of this book. 

I would mention that it can one positive effect: due to potential legal problem Linus Torvalds was promptly shipped from Transmeta to Open Source Development Labs (OSDL), the cooperative financed by several big companies (Intel, IBM, HP) interested in Linux. Unlike Transmeta OSLD looks like a place where he logically belongs.

Last modified: August 27, 2008