The Future of Open Source

GNU/Linux and Free/Libre, and Open Source software represent a socially and technically significant advance in software development. Its principles and paradigms will be the predominant influence in computing – both in advocacy or use and in reaction or opposition – for the rest of this generation.

This essay offers a precis of the trends and principles leading up to Open Source, then lists and discusses the nature and significance of several adversities. Following this, it reviews the strengths of Open Source and then comments briefly on trends, especially in health care.

To begin, I'll say boldly that the future of Free / Libre / Open Source and GNU/Linux has already arrived.

By this I don't mean that change will not occur, or that open source has had its day and will pass, and I certainly don't mean that it's fully evolved. I mean what meteorologists have long known: the most “accurate” weather forecast is to say that tomorrow's weather will be like today's. This is not trivial: it means, among other things, that we must understand “today” well in order to understand what “tomorrow” is likely to be.

We are in the midst of a technological revolution in which computers have gone from being the esoteric and expensive tools of government and great corporations to being a ubiquitous tool used daily by everyone in many ways. The typical person in this society owns and uses several computers, some unconsciously. There are usually several in their auto; plus their cell phone, their pocket calendar and address book, their wall thermostat, their local telephone exchange, and possibly several in their kitchen, including their rice cooker, coffeemaker, and stove. Oh, yes, and a personal computer connected to the Internet at home and, for folks with desk jobs or using databases (UPS delivery, salesmen, surveyors, pilots, policemen, etc.), another at work.

The computer has become a commodity in some areas of life, and this has created a conflict between commoditization and specialization.

This conflict is not new to the world of software. A general model of economic innovation and progress is that one begins with specialization and progresses to commoditization. A patent is granted to the inventor of a new device to ensure that the device can enjoy protection against the competition of copyists because invention requires a sacrifice of time and effort, and because we value inventors, we protect them against competition for a time in order to allow the expense of research and development to be recouped. For several reasons new inventions are initially priced high and their benefits are enjoyed at first only by the wealthy. But if the invention advances society, then commoditization is permitted so that the new product becomes inexpensive and therefore widely available.ⁱ

Sometimes the inventor realizes this and leads the commoditization of his product, dominating market share even after the patent expires; more often the inventor tries to maintain high margins, subsequently losing market share to competitors, and sometimes his business as well. For example, Xerox maintained its margins and short-term profitability when its patents matured, and it lost market share and nearly its existence as a result.

Sometimes public policy favors commoditization. For example, the Wright brothers worked diligently to defend their patent on the airplane. And since they did not have the capital to provide as much airplane-building capacity as the world desired, nor the ingenuity to design all the variations the people could imagine, the dissemination of airplanes was at first slowed by their monopoly. One factor in the commoditization of airplane technology was the Glenn Curtis was permitted to build airplanes with movable hinged ailerons. The Wrights argued correctly that even though they used wing-warping, their patent covered the movable aileron as well. But they lost their case, partly because they had failed to build an example, and partly because the public realization was growing that it was time to permit some commoditization of the airplane.

Sometimes the inventor leads commoditization, as when drug companies respond to patent expiration by selling both a generic and a branded version of their drug. As they have no development or startup costs for the generic version, they can easily price the generic version below competitors starting up production and maintain market dominance while accepting decreased short-term profitability. (Another way to maintain market dominance is to corner the market on raw materials, but that's not applicable to the software story.)

In the software industry the same evolution is occurring. It's simply wearing a different suit. As usual, specialization and proprietary software at first fueled this revolution, making it possible to have a ubiquitous “desktop” computer with wide connectivity for file interchange (information exchange, i. e., communication). But, as always, this same specialization now hinders progress.

“On the desktop,” a single commercial software vendor – Microsoft – has successfully monopolized a large block of computer functionality that has become important for societal function. Despite Microsoft's posturing, their success was not based on invention. Their every significant software idea or tool was invented elsewhere first. Their success has been based on aiming always to dominate their markets, by whatever technique might work.

Ironically, their monopoly was made possible by the reputation and marketing power of IBM, which learned the value of open hardware specifications from Apple Computer; by the power of giving ordinary citizens tools for analysis and communication that had not before existed; and by the tradition of single-vendor, single-platform software that had arisen from the commercial competition and specialization of the three prior decades.

The Microsoft Monopoly has been achieved, not by technical superiority or through creativity, but by marketing and muscle, leveraging its dominance into wealth used to purchase or intimidate competitors; or attracting competitors into disclosing algorithms by feigning interest in acquisition through “potential purchase discussions;” and more than anything else, by marketing distortions in ads and “information” offered to reporters, which causes commercial spin to be reported as fact. In sum, it has acted to destroy competitors in every possible way. This achievement has energized both acclaim and opposition: acclaim from those who admire audacity, single-minded egocentric pragmatism, and strength; and opposition from victims, from those whose values mitigate against such behavior, and those who understand how much better the world of software could be if the monopolist were limited to fair and honest competition.

Apple Computer opened the door for general availability of computers when it published hardware specifications for the Apple II. This permitted the commoditization of hardware; IBM followed its lead in manufacturing the PC with open specifications. The leaders of neither company, clearly, understood the nature of the flood that would follow opening these gates; neither company was able to abate its flow subsequently. Arguably, Apple could have “owned” the PC market if in 1983 it had opened the hardware specifications for the Macintosh and licensed the OS – instead, looking to maximize short term profitability, it retreated to the traditional closed model, and collapsed.

IBM struggled with the dominance of clones, and tried to “embrace and extend” the open hardware standards it had promulgated, with special hardware and BIOS features that would give its machines greater cachet than the clones; but most users preferred the simple functionality of word processing, spreadsheets, accounting, and games, and preferentially purchased “runs on anything” commoditization.

Microsoft very early saw the strength of powerful applications in the market, and at first wrote very good – “the best,” if you listened to their marketing – spreadsheet and word processor. But Microsoft also saw the power of the Macintosh graphical OS, and worked to create competition for it, partnering at first with IBM on OS/2. But Microsoft saw an opportunity that IBM did not: with Word and Excel gaining large market share, and with Flight Simulator and similar games, and – understanding that time was important – they didn't bother with a graphical OS but simply created a graphical shell for DOS. Then they leveraged their applications into greater dominance by failing to release Windows API's to application competitors until very late in the cycle, guaranteeing that for a crucial few months they would have no competition for functional applications in their new “OS.”

Within a short time, both WordPerfect and Lotus 1-2-3 had been destroyed as viable competition and reduced to niches in the software industry; OS/2, a fine OS, floundered and died by being chronically behind the curve in running new Microsoft apps. Apple's failure to open their hardware specs (and, to a lesser extent, their API's) guaranteed they could not come close to meet the burgeoning demand, as they priced their products for profitability, not market dominance. Meanwhile, Hewlett-Packard, did understand this market dynamic and destroyed Epson's seemingly impregnable dominance by licensing the Canon laser printer engine and pricing the resulting machine as a commodity.

Meanwhile, back at the Professional Computing Ranch, Richard Stallman was fulminating about the lack of an OS for academic teaching and research, blaming this problem on the commercial monopolists, especially AT & T, who owned UNIX, an operating system that Stallman admired and coveted. He correctly noted that the proprietary OS violated the centuries-old tradition of shared intellectual product among academicians, correctly realized that the issue centered around the ownership of intellectual property, and hoisted the proprietists on their own petard with the GNU Public License, which adroitly used copyright to enforce intellectual sharing.

Now we had an open hardware system – the PC – and the principle of an open software system.

Within this milieu, the US made the Arpanet cum Internet publicly available, and Tim Berners-Lee conceived hypertext referencing, making collaborative work among programmers easy and convenient. Linux Torvalds posted a little notice in 1991 into the text-based Internet, a few score people noticed, and the snowball began rolling down the hill.

For anyone who wants a pronouncement about whether open source will be a “success,” I'll say before going any further that open source is unequivocally a success, and that it will endure. But to say this is not to say that any particular firm using open-source will survive or grow, it is not to say that any particular software tool will be widely adopted.

GNU provided a basis for a social revolution that will continue to develop in the directions we have seen it take; open source, which it birthed, it is the center of a commercial and philosophical war that will both accelerate its development and impede its widespread acceptance.

Let me step aside for a moment to clarify: GNU is not quite Linux; GNU/Linux does not quite equal Open Source; Collaborative Development is more than open source. If you feel uncertain on any of these points, fire up your favorite search engine and and start reading.

Moreover, open source is preeminently a idea, a system of values, without which the open source movement, Linux, and GNU would not exist. There is not a “commercial model” for open source for the same reasons there is not one for “democracy.”

The conflicts in which Open Source is involved are old: the individual versus his tribe; the Vikings (trader-raiders) versus farmers and fishermen; town versus gown – the ancient conflict between academicians and merchants; at the turn of the 19^th to the 20^th century, the conflict between capitalists and the cooperative in the north central US; the conflict between socialism and libertarianism.

These all pit societal good against individual aggrandizement. Open Source principles philosophically place the good of the community at a respectable level; demand that community good be considered; demand that an individual's benefit not always be above that of the community.

The most important parallels are in Science and in Medicine: Medical knowledge not submitted to the medical community for verification is quackery. Not to publish a scientific breakthrough is considered irresponsible; to consider a finding factual before independent verification is considered intellectually weak. Another parallel that most miss is agriculture: knowledge that will improve production, preserve the land, or keep animals healthy has been for decades been regarded as properly belonging in the public domain because of the importance of food to life and the importance of land to production.

This conflict is of two world views, one which exalts individual aggrandizement and contends that the community benefits as an inevitable and unconscious effect, regarding adverse consequences of selfish actions to be “side effects;” the other world view holds that the community should not suffer from the actions of individuals. To compare the effects of these world views, consider, for example, the New Russia versus the Scandinavian countries.

Philosophically, the conflict is an extension of the intellectual traditions of the enlightenment versus the closed world view that preceded it. It is academic freedom applied to software development. This is not to say that all open source advocates are enlightenment philosophers, nor that all proprietary-source advocates are closed-minded authoritarians; but that the difference in their philosophies have this character.

Historically, community structures have been more stable, stronger, and more enduring than imperial structures. Ethnic groups in the Middle East date from tribal structures of millennial duration; compare this with the relatively more evanescent and unstable character of the nations that were defined after WWI. More dramatically, consider the tribal character of the current central African conflicts. In the 1950's, British expatriates warned that when the colonial powers left Africa, its peoples would relapse into ancient tribal conflicts; and in fact, the nations defined by the Europeans have been torn by these. The point is not that Microsoft represents one tribe and open sourcers another, but that communities defined by common culture and values are enduring.

Politically, the software wars are reminiscent of the struggle internationally since 1776 between democracy as a political way of life versus authoritarianisms of various flavors. Remember, there were no national democracies before the American; the American revolution was foremost a revolution in political thinking, one that took 200 years to fully infect Europe. Open source software's growth has been as revolutionary as democracy's and much faster, because its dissemination does not require that armies kill or that governments fall; and because its rationale of community benefit is as compelling philosophically.

The conflict between open source and closed source is “religious” in nature only in that its adherents are loyal to one model or the other based on core values (presuppositions or assumptions about how things ought to be) and world view (what is thought to be real or what is judged significant), which sets our priorities and life goals. Tim Berners-Lee obviously is driven by different values and a different world view than is Bill Gates or Paul Allen, just as Mother Theresa was driven by different values and world view than Bill Clinton (even though they claim the same God). I don't mean to suggest a moral parallel – while some open-sourcers may resemble Clinton, Gates doesn't resemble Mother Theresa much at all, and it's not just gender.

An example of how disparate world views sustain conflict is in the middle east. Consider the differences in values and world view between the Palestinians, who are either Islamic or Christian, and the Israelis (who are neither, if they're not Orthodox). They both feel a right of possession that dates from antiquity to a single territory; their disagreement involves the deepest of human presuppositions and is not merely one of strongly held personal opinion. These differences can never be modified by negotiation, and the only way that conflict can end in the region is for one side or the other to capitulate (or for both to capitulate to a third power), none of which will change anyone's mind about any basic “truths.”

The point is not that Microsoft is the Palestinians and open-sourcers are the Israelis, but that zealous and enduring conflict flows from sincere differences in “what is important” and “what is best.” Unfortunately, destructive behavior by partisans on both sides of the Palestinian-Israeli conflict has diminished outside support for both; and this is also true of the war by Microsoft on open source.

Momentum always favors the status quo. People prefer stability to change for many reasons, some of them admirable. Revolutions are always destructive; they are justified if the benefit strongly outweighs the damage. Linux and Linux tools are not revolutionary; the idea of software libre itself is the revolution – yet, as we've seen, is not truly revolutionary, because it's an outgrowth of the principles of intellectual freedom that have guided and benefited society for centuries.

In the enterprise, momentum favors the closed source, proprietary software for several reasons: personal investment in the status quo by professional consultants and IT professionals, horror of the risk and the burden of institutional change by managers, and simple habit.

Doctors and researchers are junkyard engineers, picking up whatever commercially-available pieces are at hand and seem to work. The habit of looking only for commercial software results in commercially-imprisoned data sets and structures, and incompatible applications at health care centers.

Businesses find a need for many small applications and tools for special tasks. Linux has a plethora of tools for programmers because these have been its main “customers.” Linux will not have a similar large, overlapping set of small, useful business applications until it has become the “default” platform for a significant portion of an industry.

Right now, because Linux is a new platform, this is lacking, and the lack is a hindrance to acceptance by some potential users, who find that The Other Platform has a larger inventory of ready-made business tools. This is a classic bootstrapping phenomenon; a large inventory is a sign of dominance which the Linux platform has not yet achieved.

Linux has only about 2% of the desktop market, and hardly any home users, thus open source has made a beachhead but no inroads into the “home” market; this would be actually of little importance except that strong influences for adoption of Microsoft platforms has been the familiarity of executives with Windows at home and on laptops, and the analytic power of spreadsheets, which did not flow out of the dedicated IT universe.

One reason, if anyone remembers the Adam Computer, is that there's a cultural sense that “inexpensive” is “cheap, tawdry, fragile.” As much as home users find Windows incomprehensible and unreliable, they typically blame themselves.

The young lady wails, “It crashed and I lost my term paper!” But she blames herself nor not making a second copy, for not printing paper drafts; she is too ignorant to realize that the operating system itself is of an inherently unreliable design because the vendor had other priorities, higher to itself than good design. She does not know that twenty years ago Mark of the Unicorn used a crash-proof swap file; that operating systems do not have to be fragile.

There are only two major “home” software vendors left: Quicken and Adobe. If both were to aggressively port their software to Linux, home use of Linux would promptly increase incrementally. Interestingly, in Europe, the standard is that tax software is supplied by the government as in Finland, Denmark, and Germany.

The other chief use of home computers dependent on ported software is game-playing. This is an important aspect of home use, and is the main reason, I believe, that the home Linux boxes for sale are Lindows boxes.

Internet access does not at all depend on Windows, and in fact is much safer without it. To the extent that a growing number of adults and children use their home computers mostly for Internet access, to the extent that they become aware that their susceptibility to viruses and the risk of cracker invasion and use of their system is chiefly due to the platform, home use of Linux will increase. But we have no marketing clout.

Another factor is that people become accustomed to the peculiarities of the dominant applications; these special characteristics are, on sober reflection, not at all necessary for functionality or even best use. But reviewers often act as though they are: if the competing product does not duplicate the functionality of the dominant one, right down to the warts, then it's said to be “incomplete.” Windows' faults and poor design decisions thus become benchmarks for the ignorant.

Another problem is that home computing is a commoditized business. Entry into a commoditized market with a complex product such as a PC requires enormous capital and high risk for small return. In order to do this, cheap, capable, and reliable hardware is necessary, plus a “Hand-holding” installation process, multimedia capability, and games are required. It is not clear to me whether the disinterest in or inability of vendors to port to Linux is due to lack of technical expertise or to Microsoft threats.

However, Microsoft has announced that Office 11 will not run on Windows 95/98. This probably represents an opportunity for some vendor to begin producing commodity systems with Internet access, multimedia capability, some Windows games, digital photography, word processing, financial software, and spreadsheets. Whether this would provide a satisfactory return in what is a fairly mature market is greatly uncertain. I think it is possible, but I don't have the quarter-billion dollars that would be required to make a good run at it (with the help of discount stores).

In any case, Lindows systems are available at Wal-Mart for $200 to $500; unfortunately the low price creates an expectation of poor quality in the average, clueless, person, as I have found by informing people of these systems.

Ignorance is a danger to open source, and benefits the monopolist. At one time the ignorance that open source exists was widespread, but there continues to be ignorance about the reliability of the OS, availability of training and support, availability of tools, of the power and flexibility of collaborative development, etc. This is in large part due to the aggressive misinformation campaign that Microsoft has waged in the technical and lay media. Ignorance of Open Source strengths of course delays appropriate acceptance and gains traction for the Monopolist's vendor-lock goals.

Managers are like reporters: they think they run the world, but unfortunately they have no expertise in any of the important technical areas that actually make the world work. Many recognize this, and function well by consulting experts and underlings as necessary. Many do not recognize this, and trudge along unaware of powerful tools and technology that could make their enterprises more efficient.

Unfortunately, managerial ignorance of the power, robustness, reliability, and superior manageability of open source software and familiarity with the Unreliable Desktop, combined with a steady flow of falsity from Microsoft in the media, have meant that Open Source has had slow growth in industry. Open Source had adequate tools for most ordinary business needs five years ago; now there is no essential desktop functionality unavailable within the OS community.

The growth of open source on the desktop is, I judge, inevitable because of the attractiveness of the Linux platform. The rate of its entrance into non-technological firms will depend on the information about it getting to non-technical users, who are kept from efficiency and reliability by ignorance of which they are unaware.

To the extent that the future of open source depends on eloquent individual leadership, there is a risk that discouragement could slow progress. The eventual success of a movement does not depend on progressing at a rate that encourages its prime movers. And the success of a paradigm does not ensure the success of businesses based upon it. Such “failures” will be discouraging to the individuals involved, potentially vitiating their involvement as leaders.

A parallel is the development of new medical devices. The paradigm is for improvement in medical care of individuals through application of medical and engineering discoveries to new therapies and diagnostic techniques. But multiple excellent efforts “fail” for each market success. For example, currently there are several simultaneous efforts to provide inhaled, transdermal, or orally absorbed insulin effectively. All are ingenious, worthy, well-conceived efforts. Several have attracted major financial investment. But it is obvious that not all can achieve market success even if all prove save and feasible; and as we all have seen in various ways, the best quality product is not always the best market success.

Uniqueness of product, unique key knowledge, have always been the “competitive advantage” in commerce. Competitors hate commoditization of their product: they know what's happened to the family farm. Commercial success is seen to depend on achieving uniqueness in some way. Commodity vendors strive in many ways to de-commoditize their services or products.

At the other end of the teeter-totter, monopoly not only ensures commercial security for an enterprise, it permits abuse of the consumer and of other producers who seek to enter the monopolist's market. The abuses that have occurred historically led more than 100 years ago to the Sherman Antitrust Act, which has been enforced in the public interest in various ways through the decades.

The consumer issue with Microsoft is not that it is a monopoly, but that it has used its monopoly position to overprice its products, especially to enterprises; and has neglected quality and security in order to preserve its monopoly position. The competitive issues with Microsoft are that it has been deliberately destructive of competitors in several ways, beyond the forces of free-market competition. The public issue with Microsoft's conduct is that dependence on a monopolist for essential public services puts the firm effectively in control of government – this situation is a power struggle that the government cannot in the long run tolerate and which the monopolist in the long run will not win.

In September 2002 twenty six nations were reported to be considering legislation similar to Peru's. Whether this report was exactly correct is less important that the significance of widespread consideration of such legislation.

Complex enterprises, as well, find themselves held hostage by this monopoly power, and have, in Microsoft's revised licensing fees and agreements of the past two years, found the ransom high: extortive fees, mandatory upgrades not at the enterprises' discretion, unfixable security defects, complex support requirements, high equipment and energy costs. But the even higher price of freedom (software conversion) keeps them captive. Thus does the monopolistic tail wag the enterprise dog.

It is a public benefit of monopoly power that the monopolist's standards become de facto public standards; the alternative is a group of competitive incompatible, non-interoperable tools and databases. This is the situation that afflicts medical care today: we are served by a large group of independent vendors; despite the Herculean efforts of the HL-7 organization, only small strides have been made toward free interchange of medical data between applications (never mind exchange between providers, to coordinate the care of individuals).

But once public standards are created, no matter by what means, they belong to the public because they serve the public good; and public good always overrides private, pecuniary considerations. This does not mean that the monopolist must cease its monopoly, but that once its standards are adopted for public use, it must consider the public good above its own in regard to their extension or alteration.

It is especially pernicious for the monopolist to create proprietary extensions of hitherto public standards, for this involves a deception: the appearance of loyalty to the public good, while (invisibly to the non-technical consumer) creating tools and applications that are in fact incompatible with those which actually adhere to the standards. This practice is simply a highjacking of public intellectual property for the monopolist's sole benefit. Such highjacking is the fundamental public threat of Microsoft's dot-Net strategy.

Thus, through proprietary extensions of current standards (because many people ignorantly adopt these by purchasing proprietary software that includes them, by adopting the functionality that these permit) customers are surprised to find themselves trapped in an unintended vendor lock.

The use of proprietary formats and standards in government programs and documents is more pernicious still, for the result is that not only the government but also everyone who must exchange information, documents, or data with the government must also adopt the same standards. This is exactly what has happened with widespread adoption of Outlook, Word, and Excel by government bureaucracies: firms and individuals who would choose not to be Microsoft customers are forced to do so in order to communicate with or to conduct business with the government: the government gets the software at a discount because the monopolist understands that all the government's customers and contractors will have to purchase its wares – to whom no discount is offered!

We cannot expect meaningful help against Microsoft's deceptive marketing or predatory competitive behavior from government regulatory administrative actions or through judicial decisions. The government has a statutory responsibility to ensure “fair trade” but the currently dominant political-economic philosophy is that essentially all actions of commercial firms are “market actions” and that market regulation should not take into account any moral (sociological) principles, should not provide any financial remedy beyond what assuredly would have occurred in case the defendant had acted legally, and with anticompetitive behaviors, should merely re-create what looks to be an opening for competition. (I hope the Chicago School does not send a hit man after me for this blunt characterization.)

Microsoft “wore down” the federal government in the original antitrust trial, through its deeper pockets and ability to prolong the process indefinitely. The government has very finite personnel and financial resources. Then, Thomas Penfield Jackson made the “mistake” of making frank moral judgments about Microsoft in his November, 1999, decision, which an appeals court found prejudicial, and the case was reassigned by lot to Judge Colleen Kollar-Kotelly, whose special interests had been in mental health law and art, not business or antitrust law, and who certainly had no depth of knowledge of the technical field involved in this case.

On October 31^st, 2002, she handed down a decision that eviscerated Judge Jackson's decision and that in essence ratified a weak agreement Microsoft had negotiated with the Bush administration, saying, “...the parties have reached a settlement which comports with the public interest.”

For a complete index of documents on file at the US Department of Justice for this case, see http://www.usdoj.gov/atr/cases/ms_index.htm. The (Third) Final Agreement between the US DOJ and Microsoft that Judge Kollar-Kotelly approved is at

This settlement requires Microsoft to undertake some specific, limited changes in response to its specific faults that brought the case to court; but this settlement does not weaken Microsoft's ability to pursue predation, nor does it create a climate in which competition with Microsoft is any easier. It is difficult for me to envision how effective enforcement of this can occur when the supervisory technical board comprises one Microsoft nominee, one plaintiff nominee, and one independent nominee acceptable to Microsoft. The compliance officer will be a Microsoft employee, giving the officer undefined access and privileges and putting the burden of this officer's expenses onto Microsoft while exposing the officer to the risk of employee-employer conflict as well as enforcement conflict.

(News reports November 5-12, 2002 state that Microsoft has created a 3-person Compliance Committee by naming 3 of its own board members; this is not what is specified by the Final Judgment; I have not been able to resolve this disparity.)

The background of this decision and the reason for being pessimistic about future government intervention in this arena is that federal judges are being chosen for their non-interventionist philosophies toward business practices. Recently Business Week made a point of highlighting the faux education offered to judges in the form of expense-paid junkets featuring seminars. For more information on the conservative strategy for neutralization of antitrust law, see

This well-funded program to subvert public justice in favor of unrestrained business practices is justified by eloquent, self-serving people: “Business must accept–indeed, embrace–the fact that the Judiciary must be lobbied as intensively as the Executive or the Legislature if business interests are to be given fair consideration in the adjudicative process.” – Leslie Cheek, Senior Vice-President, Crum & Forster Insurance.

It has always amazed me that some rich guy can start a professional sports franchise, pay exorbitant salaries to spoiled children to do intellectually unchallenging work, charge high ticket prices, and through public relations skills and by stroking media types, create a sense of public ownership of the team. Fans' ego gratification, self-esteem, and mood come to be related to the success of a commercial venture that in reality has nothing whatever to do with themselves. Why should I feel bad about myself and my town because of a competitive loss by a bunch of basket-weaving baccalaureates earning an average of $2 million a year who work for a rich carpet-bagger whose business happens to be named after my town, the Mora Muggers?

The apogee (or maybe the perigee...) of this was when the TV industry succeeded in making the Dallas Cowboys “America's Team.” This lasted only as long as they were perennial champions and has thankfully evaporated.

Meanwhile, Microsoft's skillful public relations workers have diligently worked to create the illusion that everything they have done is for the benefit of Americans in general, that work and email and games simply go better with Microsoft, America's Software Team. Many people seem to have a proprietary feeling about the company simply from using their software. This is understandable when we think about studies of laboratory animals which have shown that the rat presses the bar most reliably when food drops only intermittently – and when occasionally, instead of a pellet, there's an electric shock (the rattus norvegicus blue screen of death).

This bizarre sense of loyalty to the hand that slaps and feeds will be a hindrance to open source progress until rattus consumericus discovers that there's a kinder hand at the end of the other maze.

Less facetiously, the ubiquity of Windows-based tools and software is an important factor in perpetuating incompatibility and closed-ness of medical and scientific applications. Scientists and medical researchers are “junkyard engineers,” tending to pick whatever useful parts are near at hand to build their tools.

Then they're mystified that other workers (who have a fortuitously different set of parts close at hand) aren't able to pick up these tools and run with them. In medicine, every enterprise has its own home-grown or commercial application suite, with interconnectivity not a consideration, HL-7 notwithstanding. For example, I recently saw a very nice presentation of a clinical information system for diabetic-patient management by a clinician from a major university. It was based on a Win98 platform; it used a simple database of limited capability and proprietary design; it was carefully designed around the detailed personal and organizational preferences of its originators – who seemed mystified that the world had not rushed to adopt it in other centers. Meanwhile, sitting next to its main author at the seminar was an Italian physician whose center has been participating in the European effort to devise an open source clinical database for diabetes. Which system do you think has a greater long-term chance of widespread acceptance?

We in medicine make a pretense of wanting to exchange medical data, but several realities mitigate against actually doing this: First, most patients don't change providers if they can avoid it. Second, when they do, having the actual records from their previous provider is mainly a nuisance (because they're voluminous, not indexed, and organized in a way strange to the new provider) and a liability risk (because the plaintiff's attorney has infinite time to comb through the record that we only had time to leaf through).

Third, in the outpatient arena, the more services we provide, the more cash flow we generate. Having prior data abates this cash flow even as its perusal vacuums up scarce time. Fourth, the cost in personnel time (never mind money) of arranging cooperative records exchange with another enterprise far exceeds any measurable financial benefit, and any quality improvement is unmeasurable.

So why are we trying to create an exchangeable record at all? Because it's the right thing to do for the patient, and if done well can decrease errors. (And because some population predictably migrate through parts of the health care system, particularly servicemen, who with their families are shuffled around to different bases and then migrate into the VA health care system.)

In December, 2000, Steve Ballmer declared Linux and open source the main danger to Microsoft in the coming year. We saw, subsequently, continual disparagement in the IT and lay press of open sourcers and open source software.

A particular characteristic of this disparagement was that, in every area in which a Microsoft product had a deficiency, open source tools were falsely claimed to have worse deficiencies.

Microsoft has a large, skilled, and well funded marketing and PR effort that is kept carefully on message; Open Source has many individual spokesmen who put their own time and effort into its defense, but there is no OS PR division (no, slashdot does not qualify). I am not arguing that OS should form one, but simply pointing out that the Microsoft misinformation machine has been successful in confusing and misleading the ignorant and the innocent.

In November, 2002, Mr. Ballmer has been quoted in the press in the wake of the antitrust trial resolution as saying that Microsoft has experienced a moral conversion (not his words!) and from now on will be a good citizen, stopping all its prior attempts to mislead, stopping its campaign against open source, and only taking actions that are in the public good. Please excuse me for believing that the misinformation campaign has simply been taken to new heights. My mother used to quote scripture in response to this sort of phenomenon – “Satan himself transforms himself into an angel of light ...and his servants into ministers of righteousness.” Then she quote, “...whose end shall be according to their works.” Gotcha!

The point is not the Mr. Ballmer is Satanic, but that he's chronically deceptive, and the “information” given to the press by Microsoft isn't balanced or trustworthy. There is no Open Source, Inc., no Open Source Marketing Department, no highly paid Open Source lobbyists, to counterbalance Microsoft's guile.

The GNU Copyleft brilliantly uses copyright law to enforce open principles; not all open source efforts require such a restrictive license, but lesser licenses are viable partly because they are flanked by the GNU license and its restrictiveness.

In 2000, Microsoft threatened to enforce its patent rights on open source firms. This was a severe threat to all firms without the ability to mount an expensive and protracted legal defense against a firm with comparatively infinite resources, because even a frivolous lawsuit can destroy a nascent company. Microsoft clearly understands this.

In addition, there an intrinsic injustice in having to defend oneself against infringement of a secret patent, which the alleged offender cannot know he is transgressing. This whole situation has upset open sourcers, but it has had the salutary effect of motivating OS participants to license software innovations under the Open RTLinux Patent License –

The Digital Millennium Copyright Act (DMCA) is a highly restrictive modification of previous copyright law that causes several problems for open source and redefines “fair use” that hinders intellectual exchange.

The most extensive resource for understanding the issues presented by the DMCA and its children is Lawrence Lessig's site (http://lessig.org/ &

To oversimplify, Mr. Lessig argues that the DMCA is an extension of copyright that threatens fair use and that clotures sharing of cultural information by keeping significant but non-economical works from reaching the public domain.

In response to the DMCA, Alan Cox, one of the leaders of the Linux community, is boycotting America: he no longer travels to the US to give talks. Because the DMCA stipulates that it is an offense to actually describe discovered security weaknesses, Red Hat, Inc., releases description of security holes only to overseas customers.

A more troublesome legislative proposal was the SSSCA, The Security Systems Standards and Certification Act, originated by Senator Fritz Hollings (Dem.-S. Carolina), Chairman of the Senate Commerce Committee. It required that all hardware and software must incorporate certified security technology used in conjunction with copyrighted material.

This bill died in 2001 but has been resurrected as CBDTPA, the Consumer Broadband Digital Television Promotion Act, Senator Hollings' crippled child, sired by Disney and born of Hollings, with essentially equivalent restrictions.

A basic principle that has been followed in the divide between government control and entrepreneurial ownership is that services and products that meet an important public need are either owned or controlled by the public. So roads, sewer and water are publicly owned. Energy utilities, telephone, television, and radio and railroads are privately owned but tightly regulated. Medical care is privately owned with both government and private regulation of fees, capital construction, and quality.

The use of computers began with special-use functions with no interconnectivity; the ubiquitous use of desktop computers for most office functions and intellectual work created a need for connectivity that was finally satisfied with the Internet and broad-band communications.

We are now fully dependent for work-related communications on electronic interchange in government as well as in education and private firms. It is for this reason that the Internet is publicly owned with public protocols, and it must remain under public control, if only because no private firm must ever hold significant control over any government function.

The issue around with conflict will swirl is the extent to which communication protocol software tools should be publicly available or publicly owned, the degree to which the systems on which they are based may use secret, proprietary code, and whether document structure or specifications or structure used for government or public communication may be proprietary.

The needs of the “entertainment industry” will inject dissonance into this issue because consumers have used the connectivity of the Internet to exchange copies of copyrighted videos and music in massive numbers; this is considered by many consumers to be a “need” and a “right” – but the owners of copyright do not agree, understandably. Regardless of the popularity of music and videos, no one can argue that watching any movie is a need on a level with clean water, electrical power, or medical care.

The difficulty for open software is that the laws written to protect the rights and profits of the entertainment industry also restrict the intellectual freedom of software, even when the owners of the software copyright wish to make it publicly available. MonoSoft and HollySoft have a lavishly funded lobbying effort aimed at augmenting this, claiming to advocate “intellectual property rights.” Against this are a few eloquent defenders of “fair use,” the most prominent of which is Lawrence Lessig of Stanford University.

Ironically, the backend work of all the digital imaging shops run on giant banks of Linux servers. Meanwhile the suits out front, working diligently to lock in their marks forever, are cluelessly biting the hand that feeds them.

In the long run, I believe, the importance of public availability of publicly needed resources will counterbalance efforts to restrict access – but not until after highly restrictive legislation has been in place long enough for the general public and the politicians themselves to become uncomfortable.

A succinct summary of Microsoft's .NET strategy is at Mono's web site (http://www.go-mono.com/rationale.html). It says in part, “...it is a company wide effort that ranges from development tools to end user applications. .NET is a branding formative that has been applied to:

It appears that the dot-Net development platform is a seductive tool. A programmer in my institution called it a “wonderful programming environment” and indicated that if he were sentenced to work with it the rest of his career, he'd be quite happy.

And it seems that it is very competent as a web server. A page at http://www.linuxsucks.com contains this characteristically overstatement:

“Microsoft's .NET platform, which strongly promotes the use of XML, blows Unix and all its applications away as a web server. Yes, we know .NET is a combination of many technologies that already exist like Zope, Enhydra, PHP and JSP! But it's much better, and only if you are a programmer can you appreciate this. It will blow everything else away and then lead the pack because of widespread industry support and the big bad beast Microsoft standing behind it.”

The point of including this is not that it's completely true, but that the dot-Net

I would expect that the chief public risk to widespread adoption of .NET is that users will discover, after thoroughly integrating this platform into their enterprise, that they have become financial captives to Microsoft, just as current licensees of Microsoft products are discovering, and that its proprietary standards have become de facto public standards.

Microsoft will try to insert proprietary API's, protocols, and standards into the public “work space” in order to guarantee a proprietary stable market.

Those who sacrifice liberty for comfort and prosperity will soon find they have neither.

Two of the greatest strengths of the open source movement are the concepts of shared code and collaborative development. Microsoft has noticed this, and in 2001 announced the “shared source” license; they have – in their usual pattern of “innovation,” begun

community forums, integrated MSN Messenger into Web Matrix, and introduced an integrated development interface.

Microsoft's Collaborative Technology initiative, a la SourceForge, includes got-dot-NET workspaces: hosted source control, bug tracking, and community discussion boards.

Imitation is the sincerest form of flattery; anything that makes Microsoft development platforms more attractive will slow the growth of Open Source.

Open Source has been accused of intellectual property theft, particularly by Microsoft; primarily, I believe, for the offense of implementing equivalent functionality rather than for lifting proprietary code. This is not true; it's impossible to hide theft in publicly available code. I believe, however, that the reverse – inserting open code into closed source – is common.

I believe that Linux applications are important to closed-source vendors as a source of ideas and paradigms. I first realized this last summer when I read a comment by an Adobe executive that they aren't porting Photoshop to Linux “because of the GIMP.” I first thought, “how nice!” and then realized that the GIMP might be for them an intellectual benchmark and a source of ideas and paradigms.

In this regard, an important benefit of closing source is that it hides copyright and copyleft theft very completely.

Recently I voiced these suspicions (based only on my understanding of human nature) to an executive in the open source community. He replied that over beer one evening, some IT company executives acknowledged to him that they actually had no objections, in principle, to opening their code – but they could not because there was so much stolen code in their applications.

The concept of a “system” is a recent one; the use of sophisticated organizational techniques, engineering complexity, and the power and extensibility of computers has permitted the construction of large, highly complex machines, organizations, and information systems. These are so complex that they acquire second and third order properties that were unanticipated by their creators. For example, one of the truths discovered about systems is that they fail in ways that are actually not predictable from simply understanding their components and structure. The ValuJet crash in the Everglades a few years ago, caused by the loading of an oxygen generator among flammable cargo, has been used as an example of this phenomenon in an analysis, The lessons of ValuJet 592, by William Langewiesche published in The Atlantic Monthly, March 1998 (available via http://pqasb.pqarchiver.com/theatlantic/).

One of the dangers to a large business, a government, or the military, is that once a system is built by a commercial vendor, there is essentially no competition for its replacement. The costs of replacing an entire system are so great – capital investment, training expense, conversion costs, business interruption, and so on – that replacement is almost an unthinkable burden financially and managerially.

For the public and the government (including the military), the risk is that Microsoft will continue to insert proprietary API's, protocols, and standards into the public “work space” in order to guarantee a proprietary stable market, gaining control over functions of its own government, which it could hold hostage for for political, regulatory, market, or financial ransom.

The International Institute of Infonomics, University of Maastricht, The Netherlands, has famously issued a report reviewing the European use of Free/Libre and Open Source software, and Stanford University has begun another with respect to the US.

The Mitre Corporation in 2000 issued Open Source for Military Systems (may not be available) and in July 2001, A Business Case Study of Open Source Software (4 KB .pdf) by Carolyn A. Kenwood, outlining the ways in which open source software could be used in the Department of Defense, and on January 2, 2003, issued a report (4 MB .pdf) recommending that the military make extensive use of Free and Open Source software. These are thorough and lucid papers.

The widespread impression that Windows is The Future has led many colleges to offer only Windows-based IT instruction, making it more difficult for Unix and Linux users to recruit and train skilled workers. This will change if colleges can be helped to realize that Windows is not at all ubiquitous. For a thoughtful, well documented analysis of this, read The Case for Linux in Universities (http://www.kegel.com/linux/edu/case.html) by Dan Kegel.

The “strengths” of open source are those characteristics that contribute to acceptance, quality, or growth in the use of open source software. Open Source has many important strengths that will ensure its continued use and further growth, whether or not it actually achieves “complete world domination.”

Some people disagree with or dislike Richard Stallman; but no one can doubt that his energy and commitment has been the most important catalyst for the development of open source. The single greatest strength of the open source movement is that it has harnessed intellectual freedom for technological progress.ⁱⁱ

Nevertheless, Richard Stallman did not invent social responsibility or intellectual freedom; his genius was simply to apply these ancient traditions to software development and use.

The strength of open source is social strength, the strength of a team that will accept anyone who wishes to join. Against this is arrayed economic power and entrepreneurial outlook. These are not inherently diametrical philosophies; in fact the only conflicting principle is that of secrecy: market success does not at all require secrecy; but the situation is that the enterprise possessing unique knowledge has a leg up on the competition. The belief that secret or unique knowledge reduces risk motivates the capitalist to keep code secret even against the significant economic benefits of shared research and development and dispersed quality control.

The collaborative development model and the tools which have evolved to serve it such as the Red Hat Collaboration and Management System (http://ccm.redhat.com/), SourceForge (http://sourceforge.net/ ), and CollabNet (http://www.collab.net/) are important tools that permit assembly of groups of geographically dispersed programmers with common interests. The fact that there are many such communities of various sizes ,of developers gathered around projects of common interest, means that this model will endure regardless of the commercial success of any specific project.

Putting software into the open for peers' use enforces high standards for coding, documentation, and communication. Part of this is pride: who wants to put bad work in front of peers? Part this is pragmatic: if you want others to use your work, make it easy for them to read and understand it.

These high standards will continue to attract participation in and use of open source.

It does appear to me, from out here in the center field bleachers, that open source has superior technology in some areas. It's simply better designed. This is partly because of the liberty the programmers have to devote themselves to good code rather than to the promises or the ignorance of any marketing department. But mostly it's from the redundant, distributed debugging that occurs when development is open and code is freedThis is the subject of a formal study by Reasoning, the Linux Inspection Report, which looked at the TCP/IP stack of 2 Unix variants and "another," general-purpose non-Unix OS, and 2 embedded systems, compared with Linux. It found 0.1 defects/1000 lines of code, versus .6 to .7 defects/1000 lines of code in the other general-purpose code. A summary is available from CNET, and the report itself is available in .pdf format by request from Reasoning.

“...Red Hat's Tux 2.0 Web server running on a Linux 2.4 kernel has taken performance far beyond what was previously possible....on the SPECWeb 99 benchmark) a test performed at eWeek Labs found that Tux was able to perform nearly three times faster than current Web server mainstay Apache (12,792 transactions per second vs. 4,602 tps) when running a mix of dynamic and static Web content. ...The fact that Tux 2.0 was also significantly faster than Windows 2000's Internet Information Server 5.0 Web server (5,137 requests per second) clearly shows the advantages of Tux's new design over that of a well-established Web server. The next version of IIS (which ships with Microsoft Corp.'s Whistler project) uses several ideas introduced by Tux, including the kernel-space design.”

Red Hat has developed clustering technology for Oracle that is one of the most reliable platforms developed for databases. Oracle's management said of the Red Hat technical team, “This is the best group of programmers we've ever collaborated with.” And the technology is so robust that Oracle is now using it for their enterprise platform. For an outline of its status, see http://siliconvalley.internet.com/news/article.php/1276851. The Oracle system is based on Dell PowerEdge Servers and Dell EMC and PowerVault storage systems, Release 2 of Oracle9i Database with Real Application Clusters and the Red Hat Linux Advanced Server operating system. Larry Ellison, during his keynote speech at the 2002 LinuxWorld Expo & Conference in San Francisco –

http://www.computerworld.com/softwaretopics/os/linux/story/0,10801,73503,00.html – “said ...that he supports Linux because it's 'cheaper, faster and more reliable than any other environment around. Linux is making its way into the enterprise, and more and more companies are going to be relying on it for processing mission-critical data,' Ellison said. 'I don't think we've had a single technology take off as rapidly as our clustering on Linux.' Internally at Oracle, 'we're moving aggressively, not just to jump on the Linux hype bandwagon, but we're actually using Linux to run our own business,' he said. Oracle is practicing what it preaches, according to Ellison. 'By the end of this calendar year, literally all of Oracle's midtier [internal production] machines will be running Linux' for applications such as payroll, accounting, customer relationship management, sales force automation, marketing and human resources, he said....

He showed a slide that pointed out that a 32-processor Linux cluster running Intel machines costs about $350,000, while an IBM mainframe with comparable performance costs $14.8 million.”

The FAA began using this clustering promptly. “The Federal Aviation Administration (FAA) recently deployed an early-release two-node cluster built on Unbreakable Linux technology at its headquarters in Virginia. (http://www.nwfusion.com/news/2002/0617infra.html) The cluster is used to support air traffic control applications such as a centralized information logging system. The cluster also runs a capacity management application and a program for distributing severe weather rerouting information. One of the main goals of the Linux initiative at the FAA is to do more with less, says John Kelly, project manager and senior database administrator for Kenrob and Associates, a Virginia technology contract for the FAA.”

Since then Hewlett-Packard has joined this club (http://www.hp.com/solutions1/oracle/news_events/press/linux_cluster_benchmark.html), saying on September 16, 2002, “...An 8-node cluster of HP ProLiant DL580 servers using Intel Pentium III Xeon processors achieved 138,362.03 tpmC (transactions per minute) at a cost of $17.21tpmC with Red Hat Linux Advanced Server.

A 64-bit port has been available since January 2002. (http://www.apac.redhat.com/news/article/196.html) – “...Red Hat Linux 7.2 for the Itanium Processor features the Red Hat Linux 7.2 operating system with the 2.4 kernel and ext3 journaling file system. This configuration supports the very large address space of Itanium-based systems and scales efficiently to eight or more Itanium processors running as a single system image.” And Cray has recently released a machine with a 128-bit port.

The safety of systems has been greatly enhanced with the addition of a journaling filesystem, predominantly (I believe) the work of Stephen Tweedie. Officially, “The extended filesystem (ext fs), second extended filesystem (ext2fs) and third extended filesystem (ext3fs) were designed and implemented on Linux by Rémy Card, Laboratoire MASI--Institut Blaise Pascal, Theodore Ts'o, Massachusetts Institute of Technology, and Stephen Tweedie, University of Edinburgh...”

http://www.mvista.com/dswp/embedded_revolution.pdf: “Embedded Linux is revolutionizing the embedded systems software market. The advantages Linux offers relative to traditional proprietary embedded software solutions are tremendous. These advantages translate to significant savings in total cost of ownership and time to market for developers and OEMs, and increased ability to highly differentiate products. The result is a dramatic shift to Linux as the strategic embedded operating system platform of choice in industries of all types. This shift is occurring today and is accelerating as the technology sector initiates a new business growth cycle. The competitive advantages of embedded Linux significant that even companies satisfied with their proprietary solutions will be required to make this shift to remain competitive. The result will be the end of the traditional fragmentation of the embedded operating system(OS) industry. Embedded Linux will grow to be the dominant OS solution with a majority market share.”

Fundamentally, the impetus for Linux development and all other open source development is that someone perceived a need, and also believed that this need could be met best (or only) through open source development. Tools and the Linux OS were developed to the extent that others than the original author perceived a related need and joined collaboratively in the project.

Similarly, the success in the market of open source products will depend on customers perceiving that they have particular needs – which they discover through information (their own research or marketing by open source companies) are satisfactory solutions to these needs.

Just as legislation, regulation and enforcement, and litigation present threats and challenges to open source, they can bring support as well. In the future it may be necessary to mount copyright (left) defense as hidden violations come to light or in response to direct legal challenge.

Similarly, patent filings by open source can usefully stake out intellectual territory which can be defended through litigation. That most open source tools have multiple authors or originators is a strength for litigation; for example, when an opportunist copyrighted the “Linux” name, hundreds of lawsuits were filed simultaneously against him in hundreds of jurisdictions, leading to a very quick capitulation.

In the wake of the Peruvian open source legislative initiative, the community has become more politically active. This is important, because the closed-sourcers are spending lavishly to influence legislation. Our community does not have the lucre, but we have community interest on our side.

Importantly, Judge Kollar-Kotelly provided in her opinion a basis, to which we can appeal in our lobbying, “the public benefit” as a benchmark against which to judge the effects of proposed legislation or regulations.

Marketing is the process of providing information to people who make purchase decisions. If the information is false, then the purchase will feel betrayed – only of the information is determined to be false. The open source community will be strengthened by providing clueful, truthful, balanced information to the media; by this means we will regain respect that crackers have lost.

It's important for long-term success to provide information, not spin. Eventually people figure out that they've been hoodwinked, and when we've lost their trust, we've fractured an alliance.

The fight against commoditization by closed-sourcers is a losing one, for as the public discovers a “need” for any new product or service, growing pressure evolves to make it widely available through commoditization. Open source is already commoditized, making it easy to offer to meet community needs.

The secret to commercial success – enduring income – is secret benefits, secret knowledge, uniqueness; but as in medicine, the way for open source businesses to take advantage of this fact is to provide services and customizations that uniquely meet the needs of customers, not by putting them in a straight jacket.

For example, amazon.com became profitable after it converted from SUN Unix to Red Hat Linux. A friend said to me three years ago, “Amazon.com won't survive until it stops selling dollar bills for ninety five cents.” They did this by converting from Sun Unix to Red Hat Linux, escaping from under the princely Sun license fees. Amazon.com has been profitable ever since, and is beginning to look strong.

In consequence of the recession and competition from Linux vendors, Sun, which had a 46% margin on gross income in 2000, in the fall of 2002 it laid off about 1100 workers, and more layoffs are expected. It is not profitable right now, and analysts' forecasts call for a return to small profitability in 2004 at the earliest.

In comparison, Red Hat's gross revenues grew 30% in the past year, with gross margins are over 50%. This doesn't mean that Red Hat is getting wealthy; it means that its open source service business is seen as cost-effective in the current economy.

Most commercial banks are converting their Unix systems to Linux. Two years ago, when told they could save $100M with Linux, they would say, “We'll be interested when you start talking about real money. Since the tech bubble burst, the $100M seems to them a worthy savings.

Linux systems are just beginning to enter the retail POS (point of service) sector: this is currently 69% windows, 2% Linux: Papa John's, Regal Entertainment (theaters), etc.

Costs favor Linux/open source in terms of absence of licensing fees, control of development and development costs, improved reliability, control of security, and ability to use old hardware, lowering capital costs.

The successful commercialization of Linux is an indication that it will continue to be important. There are various models along a continuum: Debian is barely commercial; Red Hat has positioned itself as a custom programming house and service business and is prospering; IBM is staking its future on Linux and on centralized management of commercial databases (I think this will fail because people don't like to have their data stored only remotely and under someone else's control and subject to mining).

Recently SUN has claimed have joined the Linux parade; although ironically when they announced publicly the availability of their Linux platform, they were running Red Hat Linux, which led the Red Hat developers to believe that the Sun port was being troublesome to them.

Linux is primarily displacing Unix. The extent to which it supplants Windows will depend on whether applications vendors port their applications, on the security and stability of future Windows platforms, and on Microsoft's licensing terms and costs.

Expectations are a stronger influence on progress than is reality If at any time DOS had had the quality and functionality of GNU/Linux, it would have been even stronger. If at any time WIN 3.xx -95-98 had had the quality and functionality of X/Gnome, Linux might not have made much progress.

The future growth of GNU/Linux and of open source will be based in the perceived quality of OS, tools and applications, the suitability of these to tasks they're applied to, and users' ability to maintain systems (engineering maintenance, adaptation to new peripheral hardware, recovery from breakdowns, adaptation to new input or output requirements (new diagnoses, procedures, drugs, codes; changed 3^rd party requirements, management reporting requirements, and so on)

Its growth will also depend on users' ability to predict costs and keep them within the limits necessary to show profitability. Because purchase and deployment decisions are made on projected costs, not actual experience, the growth of Linux will be based on Expected costs in comparison with executives' expected ability to pay them (budget planning drives decisions).

Eric Raymond and others noted four years ago that because of Micro$oft's dominance, open source is the only way to compete in the commercial arena. Microsoft is not less dominant, less competitive, or less interested in being the sole player than it was then. There is no other competitive operating system in the marketplace. Linux is better designed: it was designed from the beginning with security in mind, it is easier to maintain, uses hardware and energy more efficiently, and is much lower in cost (the user only pays for service, not for licenses) and all these characteristics make it a challenging competitor for Microsoft.

It isn't possible for anyone to see the future. What I have tried to do here is to remind you of the most important forces that will determine the direction that Linux and open source will take in the future. The feasible question is not, “What will happen, but “In what direction is the river flowing? Where are it's banks shaping up?” For the Software Universe is a river, and we are riding along in our little boats, sometimes enjoying the scenery.

I believe that Linux will continue to grow because of technical superiority and because of collaborative development strengths. If I'm incorrect, it will be because user are unsuccessful in persuading applications vendors to port to this platform in order to avoid the excess costs entailed with using Microsoft's products. The current recession is likely to encourage the use of Linux for economic reasons.

Open source will continue to be useful because of the powerful collaborative model that has grown out of it. It will be particularly important for the “developing world” where companies, because of financial straits, have the choice between using pirated Microsoft programs or freely available open source software.

Remember that Judge Kollar-Kotelly held in her decision, “...the parties have reached a settlement which comports with the public interest.”

This is an opportunity for OS advocacy, for if it is in the public interest for Microsoft to remain intact and healthy because the public has come to need their products (which is essentially what the Judge says in her long opinion), then it is even more in the public interest to have products of equivalent functionality that are also in public control: Open Source.

Because there is no single enterprise representing Open Source, there is a cacophony of voices speaking for it – OS advocates have done a remarkable job of defining common themes and remaining “on topic,” through eloquent debate, and we need these and other individuals who are able to get the attention of the media to continue to speak. Debate within the OS community must continue, for this is how we sharpen ourselves; but we must continue to repeat the essential principles of public access and public control of shared software tools for the public good. Open Source is in the public interest.

For a partially clueful review of Open Source business status and prospects, it's worthwhile reviewing the Business Week special report of May 16, 2002 -- see

In “Giant Steps for a Software Upstart” it makes the common mistake of thinking that “Linux” is some kind of monolithic entity like AT & T. In this it confounds the operating system (Linux) with distributions (Red Hat, Suse, Debian, etc.) and also with Open Source, which is a set of principles, not a product. To say, “But will a real business model ever evolve?” is like asking whether there is a government model for democracy, as if the principles are somehow going to die unless they are manifested in a particular form of government.

The interest of Business Week is, of course, purely in the success of businesses using Linux distributions; the characteristic, concretely pragmatic world view of the typical business news analyst is hardly conscious of the powerful effect of the systems of ideas and values that are the foundation of the structure in which businesses function.

Business Week again featured Linux in its March 3, 2003, issue, The Linux Uprising. Although it perpetuates some of the myths about Linux, particularly that GNU/Linux software is written by primitives in garages and basements rather than the experienced professionals that actually do so. (It's been estimated that most open source developers work on applications with commercial support, not as hobbyists.)

Whether any particular business depending on a Linux distribution for its viability will thrive or endure in the market is related to the standard business challenges of availability, market opportunity, marketing, customer satisfaction, cost-effectivelness, and capital; success will not be directly related to the strength of open source as a model for business, political, and academic communities that are coming to understand the principles, and to use the strengths, of Open Source software development.

Open Source is an idea – like “democracy” or like “scientific inquiry” – that shapes the foundations of enterprise. An idea does not require a capital investment; it does not ask about cost-effectiveness; it does not have to show an operating profit: an idea defines our understanding of what is “best” and by molding our ideals it shapes our behavior and alters society.

And this is the reason why “Open Source” is and will continue to spread and be successful, even while particular applications emerge and evolve or fade into obscurity. Linux is not at all what it was in 1991; in another ten years it will have evolved into something significantly different than what it is today; in twenty, if it continues to evolve, it will probably look, to someone unaware of its history, as though it is a different OS entirely.

I believe that the ability of health care enterprises to participate administratively in open source initiatives has been delayed by three things:

Few Health care institutions have designed and written their own EMR; even fewer their own billing software. Those who have are primarily leaders, who began using computers before there were any significant vendors. In health care we have an extra layer of complexity between the user and the OS because of one or more layers of vendor applications.

My own institution has been struggling to interface a large number of disparate vendor systems in order to produce a geographically and technologically distributed software system for the Mayo enterprise. I recall a proposal a year ago, in which the savings of centralizing all purchasing and inventory was outlined. The projected savings were enormous – but the presenter admitted that his report did not include an ordering and delivery system that would provide the immediate response achievable by making local purchase decisions and local ordering.

I judge that the only “open source” software available to health care institutions now and for the indefinite future is the VISTA system. This has the strength of having been devised with programmers and end users working together, to ensure that it is technically and ergonomically sound. Apparently the Suits did not get overly involved in its design.

More typically, the people who specify and purchase software in the medical enterprise are the “Suits,” who despite having their own expertise, understand neither its technical qualities nor its ergonomic usability, and typically do not ask either technicians or employees to test it thoroughly. The result is that they often purchase promises and are surprised by the frustrations of their employees – but think they've done the best they could, because they don't really understand how to ask for expert advice or to discriminate it from fine-sounding fictions.

Each of us must speak out, to maintain the values of the community: those who receive must contribute. “Stealing” free code is OK; stealing work is not.

It's important that we not become discouraged with our failures – the lack of prompt acceptance of our work and our ideas. Many “failures” precede general acceptance of new paradigms. All we can realistically do is to keep talking, continue telling the truth, keep our oar in the water, and wait.

Footnotes

iNot all works become commoditized. Recently there's been legislation in the US to extend copyrights almost indefinitely. It could be argued that a societal justification for the Sonny Bono Copyright Extension Act and the Digital Millennium Copyright Act (DMCA) is that society has no “need” for any particular piece of music, art, or entertainment; there is no societal need for Mickey Mouse or The Hound of the Baskervilles. The implication is that commoditization of such works can be prevented indefinitely without harm to society.

The defect of such reasoning is that these works of art define our culture. Copyright protection means that works are available to the public only if they have sufficient enduring mass appeal to remain profitable. All other works vanish from the public because they are not allowed to enter the public domain because of their unprofitability, and subsequently fail to enrich the culture.

A solution to this dilemma is to permit extended copyright on only those copyrighted works of art that remain in commercial production; to sunset the copyright on any work of art that is not commercially available for a certain time, e.g., five years.

iiSome people ignorant equate “free software” with gratuity and costlessness – humorously derided in the open source community as the “free beer” mentality. The ambiguity of English obscures the fact that the main idea is software libre and absence of royalties is a useful tool for ensuring liberation.

The “free beer” mentality is simply a reflection of human greed – wanting value for nothing. Ironically, this is simply the obverse of financial exploitation of proprietary software – wanting excessive reimbursement for value – by its owners.

On both ends of this greed spectrum is a common characteristic: greed seeks disproportionate personal benefit for unjust compensation. At one end, exploitive licensing fees that endanger the suer firm financially; at the other, exploitive use of software without compensation for the developer's time, effort and expense.

Thus there are two reasons for pressure against proprietary software companies: One is the intellectual pressure to release to the community tools that have become necessary for community functioning. The other, perhaps equally significant in energizing the software libre movement, is the tradition of extraordinarily high fees for licensing and service historically charged by software companies, resulting in tremendous profit margins.

These charges have been determined largely by calculating the economic benefit of software to those users who enjoy documentable productivity gains and extending these fees to other users whose gain can't be so readily confirmed.

This made IBM enormously wealthy in the 1960's and 1970's and has made SUN enormously wealthy in the 1990's.

Open source commercial software is changing this: Unix vendors are being affected first: SUN in 2000 enjoyed a 46% profit on revenue. They have lost client after client to Red Hat Linux and to IBM running Linux, and have been slow to adopt Linux themselves. In September, 2002, SUN laid off 1100 employees, and has laid off more since then.

The current recession has focused many companies on economically more efficient software products. This has been a strong impetus toward conversion. The easiest conversions are from Unix to Linux. As Windows workalike software becomes available, conversions from Microsoft to Linux will also become easier. (I believe that the main cost of software conversion is re-training the end-users in a company on new software interfaces. This is as much a psychological and political struggle and cost as a technical one.)