THE POWER OF OPENNESS
Why Citizens, Education, Government and Business
|[I]n the world we live in, production is highly organized and efficient and commands enormous financial resources and seductive powers of persuasion, while demand is fragmented, uninformed and powerless. While consumers can still kill a product that they have no desire for, they are nearly powerless to direct or even influence the detailed designs of those products. For software products, consumers can only choose among a given set of alternatives, which are extremely complex, dauntingly impenetrable, and generally designed more for the company's anticompetitive purposes than for the user's tasks. |
The revelations of the current antitrust trial against Microsoft dramatically illustrate this point. The U.S. Department of Justice alleges that Microsoft has used its dominance of the computer operating system market to unfairly restrict consumer choice of other operating systems, browsers and other software applications. As the maker of more than 90% of all desktop computer operating systems, Microsoft has been able to charge higher prices for its software through restrictive licensing agreements with computer equipment makers, despite the availability of superior alternatives. Through a strategy described at the Microsoft antitrust trial as "embrace, extend and extinguish," the company has taken open standard protocols such as HTML (for web pages), Java (the cross-platform software), RealAudio (the Internet audio software) and QuickTime (multimedia software), and sought to sabotage them by introducing its own proprietary modifications as the de facto standard. Microsoft has also aggressively used a FUD strategy - "fear, uncertainty, doubt" - to scare would-be competitors from entering markets Microsoft wants to dominate. Such aggressive actions have chilled innovation in the software industry by forcing other developers to retreat to market niches where fair competition is possible.
While Microsoft may be a singular force in computing, its alleged tactics are simply an extreme case of what other software companies would likely use as well, if they could. Many software companies aspire to use ingenious proprietary design standards to create or dominate new market niches, "lock in" their technology, and preemptively limit competition and user choices. For the most part, the "seller side" is committed to determining what types of software innovations will be allowed to compete in a given market; what types of licensing arrangements shall govern users; what prices consumers will pay for products whose incremental costs of production are virtually nil; and what websites will be given preferential access or development by the primary Internet gatekeepers/portals.
It is important to understand that software design can be used strategically to shape and control markets and even national cultures. This is illustrated by PICS, the Platform for Internet Content Selection, a sophisticated HTML standard that allows private agencies, governments and corporations to rate and block online content. Although "content-neutral" in design (it can filter out birdwatching literature as easily as Nazi Party diatribes), the PICS technology is an unparalleled tool for censorship on a global scale. The development of PICS by a small cadre of little-known technicians and business executives suggests how software can be used to "design" the character of markets, political life, and culture. The so-called Y2K or "Year 2000" problem also illustrates how even a small but structurally entrenched problem in software design (the truncated field that bedevils storage and retrieval of the year value for years that do not begin with 19xx) can affect everything from air traffic control and international currency trading to emergency services.
A. The Importance of Source Code in Software
Introduction | Table Of Contents |
Proprietary source code, protected by copyright law, plays a central role in enabling the "architectural design power" of software. Originally, copyright arose to protect and reward creative expression in printed media in order to assure its public dissemination. But copyright is often misused as a legal tool for withholding creative expression and controlling the terms of competition in a given market. The seller can prevent users from altering the functionality of a software product, for example, or limit its interoperability with other software and hardware. The seller can prevent users from customizing the product to suit their own needs. The seller can ignore known bugs and other product flaws, and in other ways force users to accept unwanted design features in a product.
Historically, software makers have sought to preserve these prerogatives (and thus their relative advantage over competitors and consumers) by vigorously protecting their "source code," the internal blueprint of a software product that is the basis for the "binary code," or binaries, the user-accessible manifestation of a software product that a computer executes. As long as the proprietary source code can be held as a trade secret protected through copyright law, sellers retain greater control than users over the terms of software design, quality control, pace of innovation and use.
However important proprietary control of source code may be in some respects, that control diminishes consumer sovereignty, accountability and choice. In many instances, it results in anticompetitive mischief. This is especially pernicious because, in rapidly changing markets with technically complex products, neither consumers nor antitrust regulators are truly equipped to challenge such marketplace abuses except in the most egregious instances (such as the Justice Department's action against Microsoft), and the ultimate remedies secured may be deficient in any case.
The new software movement offers a radically new basis of competition, innovation and consumer power in the marketplace. Propelled as if from nowhere by converging forces, the new software movement has been catapulted into the mainstream over the past year. This user-driven model being generated by a vast global community of computing irregulars not only rivals proprietary software in quality, reliability, flexibility and price; it effectively "belongs" to the user community and therefore offers a cheap, versatile and durable scaffolding for the self-development of all sorts of user communities.
Just as the cooperative movement has enabled workers and consumers to assert greater control over their economic and personal lives, so the new software movement opens up new frontiers of self-determination. It allows users to pole-vault over the cost inefficiencies, barriers to innovation, consumer manipulation and design rigidities that characterize many proprietary software markets. (Open code communities are able to bypass most of the traditional impediments to coops because the Internet allows incremental and ad hoc participation, participation without regard to geography, avoidance of overhead and management costs, and rapid feedback loops with consumers.)
For all its promise and remarkable resourcefulness, the new software movement faces many serious challenges if it is going to develop a more substantial influence in the marketplace, in life, in cyberspace. Before examining that issue in Section II below, it is important to understand some of the deep dynamics of the new software movement, what makes its software so robust and innovative, and how it could, with the kinds of assistance H20 proposes, enable myriad social, educational and civic benefits.
B. The Growth of the New Software Movement
Introduction | Table Of Contents |
In the early days of computing, a great deal of software development was a collaborative process carried out by academics and students. Software was seen as the shared product of a community, to which everyone freely contributed and benefited. In this respect, open code software shares the strength and resiliency of the scientific method and Jeffersonian democracy. All procedures and outcomes are subject to the scrutiny of all. Openness allows error to be more rapidly identified and corrected. Innovation and improvement can be more readily embraced. Openness builds accountability into the process of change.
The commercialization of computing in the 1970s and 1980s introduced a new dynamic to software development: a closed, proprietary process that mobilized expertise to develop innovations, bring them to the market, and reward private investors. Bill Gates' entrepreneurial passion was so great he was nearly expelled from Harvard for using publicly funded labs to create commercial software - a violation of the hacker ethic of sharing (and not privatizing) community knowledge. After Gates was required to put his code in the public domain, as free software, he quit Harvard and went on to found Microsoft. The high-tech enterpreneurialism that Gates exemplifies is now the stuff of American legend.
Yet lurking in the shadow of this mighty new industry over which Bill Gates reigns, the free software movement has quietly persisted and grown. Empowered by the Internet, the community of computer aficionados willing to develop, improve and freely share software among themselves has mushroomed. Even as Silicon Valley gives birth to dozens of startup ventures each year, driven in no small part by the connectivity of the Internet, so the hacker community has grown in sophistication, breadth and international scope. It has generated hundreds of top-quality software programs, many of which have become critical operating operating components of the Internet.
What most distinguishes this free software from the off-the-shelf proprietary products is the openness of its source code - and the user's freedom to use and distribute the software in whatever ways desired. Anyone with the expertise can "look under the hood" of the software and modify the engine, change the carburetor or install turbo-chargers. Inelegant designs can be changed, noisome bugs can be fixed (or introduced). Sellers cannot coerce users into buying "bloatware" (overblown, inefficient software packages with gratuitous features), Windows-compatible applications or gratuitous software upgrades contrived through planned obsolescence. Nor are constant upgrades in computer hardware (such as the latest, high-speed Intel chips) required.
In short, open code software allows users to assert much greater control over their computing environment. This is not a matter of customizing a spell-checker; users with access to source code can change basic functionalities of their software and reap significant new efficiencies, often for free. By enabling skilled users to customize software to suit their special needs, open code software represents an entirely new kind of media empowerment. It potentially allows academic specialties, libraries, distance educators, civic organizations, business enterprises and others to develop their own innovative vehicles for sharing and elaborating a common body of knowledge. They do not have to adapt to the design structures, constraints on innovation and licensing/price schemes imposed by proprietary vendors. In this sense, open code software is not just a product, but a new kind of knowledge- and community-building infrastructure.
The benefits that users reap from open code software - customization, innovation, education, security, efficiency, reliability, cost savings - are actually "symptoms" of their collective empowerment as users. By banding together to assert their common interests, open code software users acquire an entirely new dimension of power. They are not merely consumers picking and choosing from the products sellers may choose (or decline) to offer; they are co-producers in the creation of specific software to serve their distinctive needs. The power wielded by the open-code user networks is vastly greater than that which propelled public interest initiatives in other media -- universal service in telephony, weak rights of citizen access in television and radio, a chronically beleaguered public broadcasting system; and local access channels for cable television. All these initiatives have required political interventions, court litigation or significant government spending, which means that have been constantly vulnerable. By contrast, the open code software movement is a sovereign, expansive "political" force in its own right.
C. The Alternative Economics of Open Code Software Development
Introduction | Table Of Contents |
Although critically aided by the Internet, the power of the new software movement stems from the "gift culture" that lies at the heart of the open code development model. The Internet is a quintessential example of a gift culture. People are willing to make all sorts of useful information available for free, in defiance of orthodox economic "rules" that claim such voluntary behavior can occur only with financial incentives. The Internet is so robust precisely because people are giving of themselves without demanding a specific contractual payback. This is the very essence of community and civility. People are willing to enter into gift economies because they trust that they will at some point share in the "wealth" that the community freely passes among itself - much as an academic community freely shares its knowledge among its members and disdains those who seek to financially profit from the community's shared body of knowledge. The key point is not that the information or software is economically free - or "free beer," as the jibe goes - but that it is freely accessible and amenable to modification by anyone interested in doing so.
(There are, of course, other reasons why people contribute software and open code to the public over the Internet. Software companies may want free publicity or a greater influence over user "mindshare"; a software developer may want to showcase his programming prowess; an academic community may wish to collectively share its resources.)
The phenomenon of leveraging the moral commitments of a community of people to achieve economic development and collective improvement is not exotic; it's simply denigrated by mainstream economic theory. Micro-lenders such as Grameen Bank in Bangladesh and South Shore Bank in Chicago have proven that they can make safe, profitable loans to inner-city entrepreneurs and collateral-poor mom-and-pop ventures that conventional banks find risky. How? By realizing how trust, moral commitment to one's peers and a concern for reputation are not only economically valuable but often more powerful and motivational than economic incentives alone. Similarly, the "Time Dollars" program has also shown how moral and social values can be harnessed to achieve "economic" goals. People can earn "time dollars" (one hour, one "time dollar") to babysit, rake leaves or provide legal assistance for other members of the community; the credits can then be used to "buy" services from other participants in the network. With no exchange of money, a considerable amount of economically valuable activity is performed - work that conventional markets are often incapable of eliciting.
So it is with open code software, as facilitated by the Internet. Many of the crucibles that incubate open code software are "gift economies," whose economic development is tightly linked with community-building. By the lights of mainstream economics, the moral and community dynamics that help generate open code software simply do not make sense. But in real life, forget the theories: open code software development works. But however formidable its creative capabilities, open code software will be developed only in those domains where its value is recognized. And many cash-strapped enterprises within academia, the nonprofit world and community organizations - as well as many businesses -- do not yet recognize the remarkable tools they could develop. The leadership and education that H20 proposes could make a significant difference.
Richard Stallman, a programmer at the MIT Artificial Intelligence Laboratory in the 1970s, was one of the first to recognize the tensile strength of what he called "free software," in which "free" refers to the freedom to change the source code, not free in price. For both personal and philosophical reasons, Stallman has strenuously objected to proprietary software because its copyright licensing restrictions limit his personal freedom to share programs with his friends and co-workers, to change the programs as he wishes, and to distribute improved versions that will benefit the larger community. He considers copyright control of software to be philosophically repugnant because it cripples the very advantages of digital technology -- its flexibility and ease of reproduction and sharing -- and encourages personally intrusive kinds of enforcement. As his peers trooped off to make fortunes in the emerging computer software industry in the early 1980s, Stallman instead founded the Free Software Foundation. The visionary project (which earned Stallman a MacArthur "genius" award grant) is dedicated to developing and promoting free software.
D. The GNU Project and the GPL
Introduction | Table Of Contents |
Two innovations developed by Stallman and his colleagues have special relevance to the new software movement. First, Stallman in 1983 launched the GNU Project, a collaborative endeavor to develop a free Unix-like operating system (free in price as well as open source code). ("GNU" is an acronym for "GNU's Not Unix," a recursive pun; Unix, of course, is the open standards-based operating system developed by Bell Labs in the 1970s. Unix was the original operating system of the Internet and of governments and academic institutions, a role that it continues to play to this day.) A primary motive behind the ambitious GNU project, Stallman explained, was to "bring back the cooperative spirit that prevailed in the computing community in earlier days -- to make cooperation possible again by removing the obstacles to cooperation imposed by owners of proprietary software." Ultimately, Stallman wants to supplant proprietary operating systems by making them free, open-source code commodities. By the 1990s, the GNU Project had found or written all the major components of an operating system except one, the kernel.
Fortuitously, a free kernel developed by a Finnish student, Linus Torvalds, and greatly improved by a large self-organized network of volunteers, appeared on the Internet in 1991. This kernel, when combined with the Unix utilities developed or collected by the Free Software Foundation over the past twenty years, "was like setting a match to dry leaves, creating an entirely new and completely open operating system overnight," according to one account. The conjoining of the GNU system with the so-called Linux kernel (Linus + Unix = Linux) has largely fulfilled Stallman's vision of a well-designed, free Unix-like operating system that can run on diverse hardware platforms.
There are other free operating systems that command great respect among software developers, yet none has achieved the huge popularity that GNU/Linux has achieved in recent years. It has been a "stealth" operating system because technophiles in major companies often secretly used the software, knowing that their bosses would disapprove of the use of a "free" operating system of unknown provenance in mission-critical applications. Now that the reliability, versatility and price of GNU/Linux has received mainstream validation, a juggernaut has been unleashed. It has become the fastest-growing flavor of Unix and a preferred operating system in such diverse organizations as Boeing, Northern Telecom and NASA. It is also noteworthy that many enterprises and universities in Mexico, China, France, Australia and eastern Europe - fed up with paying expensive licensing fees for less reliable, versatile proprietary products -- are adopting GNU/Linux.
GNU/Linux usage received its biggest boost in 1998 when the press discovered Linus Torvalds, making him the iconic underdog of the computer world. Meanwhile, over the past year, one major software developer after another - IBM, Oracle, Corel, Inprise, Informix, others -- has announced they will port software applications to run on GNU/Linux. A number of vendors -- Red Hat, Caldera, SuSE, Pacific Hi-Tech - have begun to sell GNU/Linux along with documentation and technical support, attracting significant investments from the likes of Intel and Oracle. It is estimated that more than seven million people now use GNU/Linux as their operating system.
GNU/Linux might never have emerged but for Stallman's second innovation: the GNU General Public License (GPL), sometimes known as "copyleft." Stallman astutely realized that simply putting free software into the public domain was not enough, because anyone could make minor changes in a program and then copyright it, converting it back into a proprietary product. Without some legal vehicle, the benefits of free software could be privatized and withheld from the community of users. So Stallman developed the GPL, which is essentially copyright protection with special contractual terms. "To copyleft a program," writes Stallman, "first we copyright it; then we add distribution terms, which are a legal instrument that gives everyone the rights to use, modify and redistribute the program's code or any program derived from it, but only if the distribution terms are unchanged. Thus, the code and the freedoms become legally inseparable. Proprietary software developers use copyright to take away the users' freedom; we use copyright to guarantee their freedom."
The GPL offers the greatest legal assurance that open source code will remain free and available to all, and that no company will appropriate the property for itself. But there are other copyright licensing variations. The BSD-style licenses -- based on the Berkeley Standard Distribution of Unix -- gives software users the option of creating derivative versions that can be copyrighted and made proprietary -- without revealing the source code. When Netscape released the source code of its Communicator browser in early 1998, it issued a Mozilla Public License, or MPL, which steers between the terms of the GPL and BSD licenses (private derivative versions can be made but any changes to source code covered by the MPL must be made freely available on the Internet).
Licenses that allow the privatization of the public pool of source code - thus creating a new proprietary line of development, often at the expense of the public pool -- is called "code forking." Stallman and other programmers are adamant about the principles of the GPL because code-forking vitiates the momentum of free software development. "Some in the open source community resent third parties taking from the public pool of software without contributing," writes publisher Tim O'Reilly in a major review of the open source movement. "In economies, this is called the free-rider problem. But despite the lack of a mandate, voluntary cooperation abounds."This philosophical divide is sometimes a contentious one, aggravated by the growing interest of technology companies in open code development models. It remains unclear if the powers of an open code "gift community" can be harnessed by software companies that answer ultimately to investors, not that community.
E. The Cathedral and the Bazaar
Introduction | Table Of Contents |
Although Stallman's GNU Project and Linux are two sentinel developments in the new software movement, they by no means define it. They are simply two of the largest manifestations of a more significant development, the open, Internet-facilitated process of software development. The idea that thousands of computer volunteers working together through the Internet might actually produce a sophisticated operating system that could out-perform proprietary software seemed ludicrous until GNU/Linux began to emerge. Open source theorist Eric Raymond explained the anomaly and thereby gave it wider credibility through a seminal online essay, "The Cathedral and the Bazaar," which first appeared in April 1997. Raymond writes:
Linux overturned much of what I thought I knew. I had been preaching the Unix gospel of small tools, rapid prototyping and evolutionary programming for years. But I also believed there was a certain critical complexity above which a more centralized, a priori approach was required. I believed that the most important software (operating systems and really large tools like Emacs) needed to be built like cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation, with no beta to be released before its time.
Linus Torvald's style of development -- release early and often, delegate everything you can, be open to the point of promiscuity -- came as a surprise. No quiet, reverent cathedral-building here -- rather the Linux community seemed to resemble a great babbling bazaar of differing agendas and approaches (aptly symbolized by the Linux archive sites, who'd take submissions from anyone) out of which a coherent and stable system could seemingly emerge only by a succession of miracles.
Raymond's great contribution was to explain in specific detail how and why a motley assemblage of thousands of hackers working for free on their own time ("the bazaar") could produce better software programs than the expensive professional talent amassed by Microsoft and other software companies ("the cathedral"). The answer has much to do with tapping into the decentralized intelligence of a huge pool of personally motivated programmers via the Internet. Others have observed that the cost of developing any software that is fundamentally better than current software will require more resources than any single corporation can afford - but that global networks of the best programmers motivated by prestige and peer recognition can generate vastly superior software, as GNU/Linux demonstrates.Some of Raymond's trenchant observations:
Treating your users as co-developers is your least-hassle route to rapid code improvement and effective debugging.
Every good work of software starts by scratching a developer's personal itch....The best hacks start out as personal solutions to the author's everyday problems, and spread because the problem turns out to be typical for a large class of users.
Often, the most striking and innovative solutions come from realizing that your concept of the problem was wrong.
Given a large enough beta-tester and co-developer base, almost every problem will be characterized quickly and the fix will be obvious to someone. Or less formally, "Given enough eyeballs, all bugs are shallow."
The analysis of the "bazaar" model of software development was so compelling that Netscape, under siege by Microsoft's Internet Explorer, decided in January 1998 to release the source code of its Communicator browser. Netscape's move was a radical notion at the time, motivated at least as much by business calculation as moral conviction. Within months, many other software companies began to recognize the power and technical superiority of open code software -- as well as the niche profit opportunities in a marketplace dominated by Microsoft. Such major players as Oracle, Corel and Informix announced that they would be porting their products to GNU/Linux. Others, such as Sun, have made the source code of their programs available in some fashion or another, but under modified licensing rules that seek to maintain some measure of proprietary control. It has been especially significant that IBM has announced its support of GNU/Linux and the Apache Group (the open code community that has made Apache the dominant web server software) and that it will ship a free, no-frills version of its database program, DB2, for GNU/Linux in spring 1999.
The growing interest of software vendors and even chip-makers like Intel in open code software promises to trigger significant upheaval in the computing/software industries as the proprietary world seeks out new ways to exploit, domesticate or co-exist with open code software. It is unclear how GNU/Linux and other open-code software will fare if proprietary developers actively seek to arrest or control its growth. On the other hand, Red Hat, Caldera, SuSE and many other companies see novel business models based on a growing and robust open code software "market."
In any case, open code software clearly represents a serious alternative to the traditional proprietary model of software development. Its appeal is not just that it is cheaper, more versatile, reliable and customizable software. Open code software represents a structural shift of power from sellers to users, and in that sense is one of the most liberating tools of media empowerment that individual citizens and the civic sector might ever imagine. Open code re-positions the terms of competition to a matrix of quality and utility, and diminishes the advantages to be gained from manipulations of the distribution apparatus, marketing schemes, restrictive licensing terms, and bundling deals with hardware makers.
Open code software is also inherently more suited to educational environments because its inner logic - the source code - can be directly manipulated by students. With its inner parts visible, users can choose to learn how the software works, and then share and develop that knowledge. Proprietary software, by contrast, is inherently "unknowable" because its inner architecture is a trade secret.
The future of the open code software movement may hinge upon what new terms of engagement emerge between the proprietary world and open code software developers. Can the integrity of the free software vision and the vitality of its online communities be maintained as the movement expands and perhaps morphs into a more diverse open source movement? Section II explores the obstacles that must be addressed, followed by an agenda that H20 proposes for fostering the continued growth of the new software movement.
II. FORCES THAT MIGHT DERAIL THE OPEN CODE VISION
Introduction | Table Of Contents |
Open code software just may be the elusive "killer app" that the computer business has sought for so long: a powerful, novel software that "changes everything." Yes, its actual presence among average computer users is still quite limited. But as the poet Mary Oliver might put it, "It's not the size, it's the surge." GNU/Linux is the only non-Microsoft operating system showing growth in market share - 212% growth in market share in 1998; it has 17.2% of the market for server operating systems, a nearly three-fold gain over the previous year.Apache, Perl and many other open code software programs are being embraced by major proprietary vendors as cornerstones for future software development. Even Microsoft concedes in internal strategic analyses leaked to Eric Raymond -- the "Halloween Memos" -- that GNU/Linux and other open code software "provide very dramatic evidence....that commercial quality can be achieved/exceeded by open source software projects."
All of this suggests that the new software movement stands at a new threshold. It has a sovereign vision, a superior product, a burgeoning cadre of supporters, growing investment, fresh attention within technology circles and the general press, and a hardy development process of proven effectiveness.
But can this movement sprout wings and fly? Open code software faces some formidable challenges if it is to grow and become a broader consumer phenomenon. Among the barriers it faces are the lack of popularly accessible documentation and technical support; the lack of a clear, well-known brand identity and marketing support; a susceptibility to code-forking that can vitiate the development process; the shortage of enterprise-oriented development tools; development uncertainties that make proprietary vendors queasy about porting applications to open-code operating systems; and problems with integrated, end-to-end ease of use. These issues raise questions about the ultimate strength and durability of the open code movement, especially if proprietary vendors seek to domesticate or divert its software in order to protect their existing products and revenue flow.
This section outlines some of the challenges that proponents of open code software may face, both affirmatively in expanding the quality and usage of their products as well as defensively in neutralizing monkey-wrenching by competitors. It is unclear which of the challenges enumerated below will actually materialize and which opportunities should have the highest priority. But if the new software movement is to flourish, these issues certainly deserve further inquiry and discussion.
A. Can the Integrity of the Open Code Vision be Maintained?
Introduction | Table Of Contents |
The dynamic tensions between the proprietary "cathedral" models of software development and the open code model are likely to intensify in coming months and years. Indeed, this very polarity may become more complicated as new proprietary/open code business models are developed. A key question is whether the open code movement can retain the integrity of its vision and community-driven development process as the proprietary world comes calling. Will established companies see open code software as a threat or promise, or both?
There is no question that Microsoft sees open code software as a threat to its franchise. As the Halloween I memo astutely notes, the most significant threat is the open software development process: "The ability of the OSS process to collect and harness the collective IQ of thousands of individuals across the Internet is simply amazing. More importantly, OSS evangelization scales with the size of the Internet much faster than our own evangelization effort appears to scale." Halloween I concedes that open code software "poses a direct, short-term revenue and platform threat to Microsoft, particularly in server space. Additionally, the intrinsic parallelism and free idea exchange in OSS [open source software] has benefits that are not replicable with our current licensing model and therefore present a long term developer mindshare threat."
Apart from such outside threats, however, there are questions about whether the open code community will retain its community vitality and mutual commitment as its popularity (and opportunities for entrepreneurial spinoffs) grows. Also, can open code communities think and act strategically? Vinod Valloppillil, the author of Halloween I, shrewdly points out:
A very sublime problem which will affect full scale consumer adoption of OSS projects is the lack of strategic direction in the OSS development cycle. While incremental improvement of the current bag of features in an OSS is very credible, future features have no organizational commitment to guarantee their development.
It is an open question whether "organizational commitment" is needed to stabilize and guide open code software development. Given the radically decentralized nature of the new software movement, however, it could benefit from a "loose leadership" that helps convene relevant players, focus discussion, scout out the frontiers and develop strategic analyses. Such efforts are important if the movement is going to respond to the inevitable threats that the proprietary software world will contrive. There may be some lessons to be learned from the 1960s, in which leader-less political initiatives were sometimes out-maneuvered by the more strategically organized, resource-rich opposition. Without losing the moral legitimacy and political strength of its grassroots base, the new software movement needs to evolve new governance/leadership structures to deliberate and advance its interests.
At the same time, the open code vision will require a broader, more ecumenical engagement with computing professionals, especially within corporate MIS departments, and non-technical constituencies such as education, government, nonprofits and the general user. Maintaining a core vision while popularizing it is a signal challenge.
B. Can the Open Code Community Overcome Intellectual Property Snares?
Introduction | Table Of Contents |
Many open code advocates correctly worry that software companies will create copyrighted derivatives that will privatize the shared intellectual capital of the open code community - the very scenario that inspired Richard Stallman to create the GNU General Public License. This has happened, for example, with SendMail, the open-code mail transport software whose original author, Eric Allman, has turned it into a proprietary product (while promising to improve and support the open-code version). This trend may accelerate as more programmers decide to become entrepreneurs feeding off of open code networks.
While such adaptations of open code programs may seem relatively benign, similar tactics by a Microsoft using "embrace, extend and extinguish" tactics could be disastrous. Microsoft speaks of trying to "fold extended functionality into commodity protocols/services and create new protocols." What this means, Eric Raymond explains, is "introducing nonstandard extensions (or entire alternative protocols) which are then saturation-marketed as standards, even though they're closed, undocumented or just specified enough to create an illusion of openness." This, of course, is precisely the Microsoft strategy that a federal court enjoined recently in Sun's litigation. It charged Microsoft with "standards pollution" of Java, which is intended to be a "write once, run anywhere" software - a "virtual machine" that can run on a variety of operating systems.
Clearly the GPL offers the best protection for expansion of open code software. But that may not be enough. Some distribution vendors combine non-GPL software with the GNU/Linux kernel and distribute them, introducing new copyright quandaries. Also Microsoft may be able to use its market power to prevent open code software from reaching a wider base of consumers. It could prohibit OEMs [original equipment manufacturers] from pre-installing GNU/Linux, for example, as it appears to be doing with Dell in Europe. Even though Microsoft has recently allowed computer makers to take new liberties in configurations of their opening screens (allegedly in response to the DOJ lawsuit), Microsoft's underlying contract terms with OEMs have not changed.
Nor are any PC makers accepting a challenge by the chairman of Be Inc. to pre-load either BeOS (his company's operating system) or GNU/Linux for free - which would have obvious value to consumers. None will do so, contends Jean-Louis Gassée, because it would violate Microsoft licensing provisions, which he says prohibit OEMs from modifying the boot manager to display non-Microsoft operating systems or allowing the display of "unapproved" icons on the desktop screen. "As an industry insider gently explained to me," said Gassée, "Microsoft abides by a very simple principle: No cracks in the wall. Otherwise, water will seep in and sooner or later the masonry will crumble." If Microsoft's licensing terms could be resisted with impunity by OEMs, on the other hand, it could allow real competition among operating systems. It would also vastly expand the potential user base for open code software and moderate the prices for Microsoft products.
There are also questions about hybrid corporate-sponsored "noospheres" (discrete communities of open code programmers) such as Mozilla, the Netscape-sponsored group of programmers dedicated to extending and customizing Netscape's browser. Now that Netscape is being bought by AOL, there are fears within the Mozilla community that AOL-Netscape could abandon Mozilla or push its work into an AOL-owned proprietary tree. AOL-Netscape, for its part, have their own fears of legal liability if the open-code process were to introduce patented code into the browser. Similarly, Sun is trying to kick-start a community of open code developers for its Solaris software using a quasi-open code license which it calls a Community Source License. The long-term viability of these experiments remains an open question.
C. Can Open Standards and Interoperability be Assured?
Introduction | Table Of Contents |
Beyond desktop and enterprise computing, the new software movement is committed to an open framework for the Internet. Fundamentally, this means open standards and interoperability among different hardware and software platforms. Open standards means that the standard-setting process is open to anyone; that technical specifications are published and freely available, and that no single vendor can control or manipulate the evolution of the standard. Ideally, there are certification tests to assure that an open technology actually meets the specifications.
The Halloween I memo suggests a chilling scenario by which Microsoft would subvert the commodity standards and infrastructure that is the foundation of the Internet by introducing proprietary protocols. The author suggests that Microsoft "de-commodize protocols and applications" by "extending these protocols and developing new protocols." This would undermine the core advantage of open code software, its ability to integrate diverse hardware and software through common commodity protocols. This is why Microsoft finds open code software in general and GNU/Linux in particular so alarming: it threatens to subvert its proprietary integration of systems - and thence its monopoly rents -- with open, competitive integration based on user-determined standards of quality and utility. This is why Microsoft quietly changed its protocols to render Samba, an excellent open code communications application, incompatible with Microsoft's Windows NT and Windows 95/98 software. This also explains why Microsoft is hostile to the IETF, the Internet Engineering Task Force, which has historically resisted the special pleadings of any single vendor in generating technical standards for the Internet.
When AT&T and the Bell Companies controlled the technical standards for telephonic network, no one else could compete. But once standardized interfaces were required to ensure interoperability, competition was possible - and innovation in all sorts of value-added products and services materialized. So it must be with the Internet, now and in the future. It is vital to ensure that the computer box, servers and switching protocols are interoperable, or else competition, innovation and their myriad benefits will be thwarted. The open code movement represents a powerful but under-mobilized constituency for assuring open standards and interoperability on the Internet.
D. Can Open Code Software Build a Brand Identity?
Introduction | Table Of Contents |
In our business-dominated culture, it is an inexorable reality that people respond to brands - an image and aura that distinguishes Brand x from Coca-Cola, McDonalds, General Motors and Microsoft. This may pose problems in expanding usage of open code software, because it has none of the contrived mystifications that come from advertising. It doesn't have an image, except perhaps its anti-image as a Birkenstock, counterculture artifact. While image may not matter much for techno-sophisticates, it could matter for the average consumer and even for the middle-managers of corporate MIS departments. Brand names are a form of cultural credibility. Even though that credibility may be unwarranted or unjustifiably expensive, it is, for now, a price of entry to the mainstream of American commerce and culture.
If it is going to expand beyond its grassroot niches and become an established player in mainstream computing, open code software may need to bring its image into focus and "manage" it with some strategic sophistication. Some argue that the modularity, technical superiority and increasing plug-and-play capacity of GNU/Linux may make traditional advertising gratuitous. This may prove true. On the other hand, the larger universe of open code software could become a perennial also-ran - an uncredentialed, second-choice alternative, despite its superior merits - unless it has some minimal branding and marketing support.
One small but important step in this direction is OpenSource.Org, a certification system started by Eric Raymond and others, which gives a seal of approval to open code software that meets ten specific standards. Under the organization's criteria, the GNU GPL, Berkeley Software Distribution, and Mozilla Public License, among others, meet the standards. (A similar open-standards certification for hardware may also be necessary as more software functionality become embedded in hardware interfaces, threatening the open protocols that allow competition and innovation.)
What really needs to occur is a more coordinated cooperative association of powerful brands. This could resemble coop ad campaigns for an entire product category -- milk, eggs, books - financed by a group of vendors. On a parallel track, there could be an aggressive public education program launched through popular magazines, the general press and constituency-specific periodicals (e.g., higher education, nonprofit trade associations). There are also online vehicles, conferences and other vernacular venues through which the open code "message" could be popularized in highly credible ways.
It is unclear whether the various new software communities might wish to cooperate and "go legit" through advertising and marketing. Certainly such initiatives would need their support. But resistance could be self-destructive because ultimately, the "counter-cultural" identities of GNU/Linux, Apache and other open code projects are vulnerable to appropriation and caricature by competitors if there is no organized, strategic plan for "managing the identity" of the new software movement. As the Halloween memos frankly admit, this is a battle for "mindshare." Part of the success of open code products will require the savvy, pro-active cultivation of mainstream awareness and acceptance.
E. Can Open Code Software Be Made User-Friendly?
Introduction | Table Of Contents |
While GNU/Linux has achieved remarkable growth among the technically proficient, its future among the average computer user will require improvements in user-friendliness. The modular architecture of GNU/Linux, which makes perfect sense for the technically proficient, may prove troublesome for the mass-market consumer. Generally speaking, open code software has not undergone extensive testing or modification with non-techies to make them easier to use. They are, after all, products of the hacker community. Microsoft has built much of its empire on the seamless integration of a user-friendly operating system and applications: a value-added monopoly function, poorly achieved, for which it charges a significant premium.
The parallels with the pre-divestiture AT&T are striking. Before the 1984 AT&T consent decree, there was no competition; just a regulated monopoly. In the years after that watershed, a flourishing telecom marketplace rife with innovation emerged. We stand at a similar pre-divestiture juncture with Microsoft, in which competition in the operating system market and other ancillary applications markets is artificially limited. A large, robust marketplace of modular, open-code applications stands ready to emerge, but has not. There is hope on the horizon, as Red Hat and other vendors begin to offer integrated packages and components, and as new desktop environments such as KDE and GNOME gain in popularity. Even more striking, IBM is now planning to offer round-the-clock tech support and service for Red Hat Linux. All these steps will help "grow" the market for GNU/Linux and open code software. But we have a long way to go before such "end-to-end" ease of use will be achieved.
F. Can Open Code Software Attract Support from Non-Technical Constituencies?
Introduction | Table Of Contents |
Ralph Nader became an early advocate of open code software because it radically empowers consumers in many ways. The benefits are not undifferentiated benefits that flow to the aggregate consumer, but highly individualized benefits that may help particular niches of consumers: education, nonprofits, voluntary civic associations, charitable endeavors, professional associations, and business users as well. As open code operating systems become more pervasive, it becomes more plausible to develop more stable, enduring open code platforms and compatible applications. The question is, can non-technical constituencies be made aware of their enormous stake in the emergence of open code software and help spur a demand-pull for it?
Since much of the actual, long-term value of open code software is not immediately understood by the layperson, this could require a lot of missionary work. Indeed, in the short term, open code software is likely to be far too technical and complex for the average user to find comfortable. The systems must evolve more. But this should not deter new initiatives to educate non-technical constituencies about how open code software represents a community- and knowledge-building infrastructure without precedent. It represents a sovereign "media space" independent of proprietary manipulations, arguably one of the most powerful potential forms of user-empowerments in electronic media.
III. STRATEGIES FOR SUPPORTING OPEN CODE SOFTWARE: