Why Open Source Works

The modern phenomenon of open source software has everybody asking questions. Namely, why would anyone give away what they could just as easily sell? It turns out that the answer to this question also answers the next: how do hundreds of people collaborate on large open source projects? The motivations that drive participation also lead to qualities that encourage successful large-scale production of software. The individual hacker (a term I use to mean amateur programmer and open source participant) feels fulfilled by doing challenging and creative work. He earns the community’s respect through collaboration, and the cooperation of all his fellow hackers ensures that the project’s goals are met. This is not unlike the happy Troglodytes of Montesquieu’s Persian Letters who worked together to succeed where their selfish cousins fought amongst themselves and failed. The cooperative advantage is plainly seen in the success of open source today, and that success is rooted in the intrinsic motivations of the individual hacker. The very things that get them to work get them to work together.

The emergence of open source communities originated with the rise of the internet. It is in the internet’s unique environment that open source flourishes. It’s one of the few places where ideas ideas and information are freely exchanged, and this very ability is key to harnessing the creative energy of hackers. However, hackers were sharing long before the internet was thought of. In 1955, a computer users group called SHARE (“The Society to Help Alleviate Redundant Effort”) was founded at IBM, only four years after the UNIVAC, the first commercially available computer, was released (Lum et al). Real hacker culture began to emerge at MIT in 1961 as students developed their own tools and slang – they were also the first to call themselves “hackers” (Raymond 8).

It wasn’t until the end of the sixties that ARPANET, the predecessor of the internet, came online (Lum et al). It was the first high-speed computer network to cross the country, and it connected isolated hacker groups around the United States (Raymond 9). Interestingly, as hackers developed a thriving culture based on sharing, Bill Gates wrote an open letter to computer hobbyists asking, “who can afford to do professional work for nothing?” The very next year, Bruce Perens wrote the first draft of “The Open Source Definition” (Lum et al).

Open source emerged on the internet because its electronic mailing lists allowed hackers to communicate and cooperate. To this day mailing lists are the major tools of communication for open source projects. Some early projects included Richard Stallman’s efforts to create high quality tools for the Unix operating system. In 1982, he began an ambitious project to produce an entire set of free Unix tools under the name GNU (for “GNU’s Not Unix”) (Raymond 17-18). However, it was not until 1991 that a full clone of Unix was produced. Using Stallman’s GNU tools, Linus Torvalds created the first version of his operating system, Linux, and submitted it to mailing lists for feedback (Lum et al). This was the first example of large-scale collaboration done via open source. “[Linux] was rather casually hacked on by huge numbers of volunteers coordinating only through the internet” (Raymond 24).

The volunteers working on Linux are just like those working on open source projects everywhere today. It is their motivations that drive the production of Linux and maintain its structural integrity despite its size and complexity. The first and most obvious of these motivations is creativity. “The programmer, like the poet, works only slightly removed from pure thought-stuff. He builds his castles in the air, from air, creating by exertion of the imagination” (Brooks 7). Computer languages are not so different from human languages, as both allow for powerful self-expression. In Lakhani and Wolf’s survey on “Why Hackers Do What They Do,” creativity was found to be the “lead driver of individual effort” (16). As artists, hackers derive a great deal of fulfillment from solving hard problems and writing great code.

Software development can become as passionate an endeavor as poetry, and just dependent on skill. “Software is intellectual property. You have to have intellects at work in order to get software” (McCarthy 30). One interesting facet of the intellectual exertion is flow. Flow is a state of mind reached during intense focus on one’s work, almost like meditation. In flow, the hacker does not perceive effort – or the passage of time or the surrounding environment (DeMarco and Lister 63). It also represents the height of creative involvement and enjoyment. If the task is too easy or too difficult, boredom or anxiety prevent that optimal state of activity called flow (Lakhani and Wolf 4). It’s not always simple for a hacker to find a problem that’s “just right,” but when he does, it is one of the most rewarding experiences of his creative life (ibid 11).

Intrinsic motivations bring people to software development, and they also encourage them to participate specifically in open source projects. Software companies enforce deadlines and production needs with strict management hierarchies and strong pressure. The traditional economic model behind these practices views human behavior as extrinsically motivated: “People change their actions because they are induced to do so by an external intervention” (Lakhani and Wolf 6). However, DeMarco and Lister found that corporate software projects on which there was no schedule pressure were the most productive in their study (29). “There is nothing more discouraging to any worker than the sense that his own motivation is inadequate and has to be ‘supplemented’ by that of his boss” (ibid 9). On open source projects, all hackers work because of their own intrinsic motivation, so there is no need for control (Brand).

However, there is still the issue of money that Bill Gates asserted. Hackers could use their skill to get a steady salary at a commercial firm if money is such a strong motivator (Lerner and Tirole 20). Yet DeMarco and Lister’s study found salary to be only weakly correlated to performance. This would indicate that money is not so strong an incentive as to actively discourage hackers from writing free software (47). Nor does it keep people from being productive on the job: Employees will work overtime to get important work done if they have to (ibid 43). It is the work itself that interests programmers, not the money. The corporate structure just gets in the way of their creative and productive work.

A more practical reason for a hacker to hack is to “scratch an itch.” If he has a problem with existing software, he can solve it by fixing it or writing his own version. The problem with conventional proprietary software is that it’s closed source: The hacker can’t access the code for modification. As the user changes and the world around him shifts, the software no longer “fits” very well (McCarthy 70). One of the inspirations for the open source movement was Stallman’s frustrations with a proprietary printer program that Xerox wouldn’t release the source code for (Lerner and Tirole 26). Apache, an open source web server that underpins much of the internet today, originated over a mailing list by developers who couldn’t add patches and improvements to an older program (ibid 13). Of course, “in many instances, solutions developed by particular users for individual problems have more general solutions for wide classes of users” (ibid 3). The motivations that drive a software developer lead to useful, quality work because that’s what he wants to produce. Open source does not provide overt extrinsic motivations or incentives, but instead steps back and lets hackers do what they do.

Perhaps it is now time to elaborate on open source itself. It has already been mentioned that open source is a cooperative effort to produce free software. A particular project is set up by one or more hackers wishing to share their efforts and produce something useful. Other members join because they want to help improve it as well. Communication is vital for this collaboration to work: Users who find problems (bugs) must report them so that developers can fix them (debugging). Developers submit their changes to fix bugs or add new features, and the community decides which patches are integrated into the next released version of the software. Some project communities vote on changes, while others rely on the original founders to make final decisions (Brand). This style of organization has evolved organically from the first hacker communities, so there is generally little extraneous structure. The community forms naturally as hackers are doing what they want to do, so its structure makes it easy for hackers to continue doing it.

Open source communities can be seen as gift cultures. Hackers can easily access everything they need, and the internet makes it easy to share their software creations. “This abundance creates a situation in which the only measure of… success is reputation among one’s peers” (Raymond 99). A hacker’s “gift” is a patch or other contribution. It’s relevancy to the community’s goals dictates how much respect it earns for its giver. The status of an individual within the community is determined by his skill and ability rather than social class or wealth, making open source meritocratic (or based on merit) (“Meritocracy”). Social standing is a strong motivation in its own right, but the simple feedback or peer-evaluation it gives hackers shapes their behavior even if they don’t play the “reputation game” (Raymond 102).

The peer-evaluation system is not only good for hackers, but for the software as well. “Continued devotion to hard, boring work (like debugging, or writing documentation) is more praiseworthy than cherrypicking the fun and easy hacks” (Raymond 117). Even though open source projects are volunteer efforts, the dirty work still gets done because hackers get rewarded for doing it. However, bragging about such work – or anything at all – is taboo. Seeking personal advantage at the expense of others weakens the collaborative power of the group (Lakhani and Wolf 5). The peer-evaluation system is disrupted when the quality of the gift doesn’t speak for itself (Raymond 109). The reputation game isn’t “won” – the respect given by other members of the community is meant to bolster a hacker’s self-esteem, not his ego. A hacker’s self-esteem is tied “strongly to the quality of the product” he produces (DeMarco and Lister 19), which is why peer-evaluation ensures quality by reassuring the hacker of his natural tendency to do well.

As a community grows, its members become more invested in the community’s well-being. Cooperation is most beneficial when it continues because it can “breed reciprocity and trust, to the benefit of all” (“Economics”). Giving away software for free is only a stretch of the imagination if there isn’t a promise of feedback and the hope that others will offer their help in return (Raymond 152). The common interest of the group is served if each individual submits to the will of the community rather than asserting his own needs. “If hackers were egotistical, rational actors with no social bindings, the collaboration would not work” (Brand). The community acts as a whole greater than the sum of its parts, with the teamwork allowing for greater productivity and greater enjoyment than the same hackers would get working separately (DeMarco and Lister 123).

Because each hacker is not acting as an individual but as a member of the group, a great number of people can be involved without a strict hierarchy. Linux is no exception: many large collaborative projects are emerging that lack managed hierarchies (Benkler 7). KDE is a graphical interface and software suite comparative to what most people interact with on Windows. Yet this impressive project is also be supported by open source collaboration without hierarchies (Brand). The reputation game does not systematically rank participants. It only recognizes those hackers who contributes most to the community. Turning around and using that power goes against the ego taboo, so leaders are forced to be humble and guide through their own contributions.

The fulfillment hackers get from being part of a successful project actually leads to its success. The creative self-expression of the hacker takes the form of software. Because his self-esteem is tied up in the quality of his work, he invests himself in his code. But “deep within, we want others to use our work and find it helpful” (Brooks 7). The peer-evaluation recognizes good work and puts it use. His self-esteem is now tied to the success of the project as a whole, because his own work is part of the greater software. Indeed, he has identified so much with the project that he and his team members have jelled to form a productive community. The community provides even greater fulfillment for the initial software creation, encouraging him to help meet the needs of the community – and both he and the community want the final product to be great.

It is useful to compare open source communities with those of the Troglodytes of the allegory by Baron de Montesquieu. These near-human creatures were originally completely self-interested, and not at all like the friendly hackers. They would lie and cheat, even kill each other, just to get the better land or the better wife (Montesquieu 1-2). However, this society destroyed itself because of the Troglodytes’ inability to cooperate in hard times. Those that did survive went on to construct a perfectly just and harmonious community. “They worked with equal solicitude in the common interest” (Montesquieu 3). The Troglodyte’s new community behaves much more like that of the hackers. The traditional software production is driven by self-interest and competition, while the open source haven has succeeded beyond all expectations. Both Troglodyte and open source communities are cooperative societies which take advantage of plentiful resources in their environments. As gift cultures, the contributions of members to community are valued above resource-hoarding and taking advantage of others: “Nature provided for their desires as abundantly as for their needs… They would give each other presents, and the giver always though that the advantage was his” (Montesquieu 4).

The mystery of open source has now been unveiled as a simple human community. Perhaps not so simple, but the trail leading from the initial question of why hackers volunteer to the final answer that their motivations for doing so also drive the success of open source has only been eight pages long. Hackers are intrinsically motivated to contribute: they love doing what they do because it’s challenging and creative work. Cooperation justifies the initial work, but also brings its own rewards. The community supports the hacker’s creations and inspire him to do more for the success of the project. Using the power of communication provided by the internet, hundreds of individuals come together to work towards a single goal, sharing resources and ideas. However, the similarities between open source communities and those of the good Troglodyte’s brings some doubt as to their future. In the end, the Troglodytes were so successful and numerous that they sought to crown one of their own as king. But the individual they asked to take power responded with despair.

“In your present state, without a ruler, it is necessary for you to be virtuous despite yourselves. Otherwise you would not continue to exist, and you would fall into the misfortunes of your ancestors… Would you want [a Troglodyte] to perform a virtuous action because I tell him to, when he would have it just the same without me, by natural inclination alone?” (Montesquieu 6)

Perhaps hackers will one day follower their leaders without question rather than with respect, or return to the salaried jobs of corporate firms. But as long as hackers love to hack, the exchange of ideas and cooperation of people seen in open source will survive.

Works Cited

• Benkler, Yochai. “Coase’s Penguin.” Yale Law Journal 112 (2002): n.pag.

• Brand, Andreas. Email Interview. By Tom Chance. “The Social Structure of Open Source Development.” News Forge. 1 Feb 2005. 14 May 2005 http://programming.newsforge.com/article.pl?sid=05/01/25/1859253.

• Brooks, Frederick P., Jr. The Mythical Man-Month. Reading: Addison-Wesley, 1995.

• DeMarco, Tom and Timothy Lister. Peopleware. New York: Dorset House, 1987.

• “Economics of Sharing, The.” The Economist. 3 Feb 2005. 14 May 2005 http://www.economist.com/finance/displayStory.cfm?story_id=3623762.

• Lakhani, Karim R. and Robert G. Wolf. “Why Hackers Do What They Do.” Perspectives on Free and Open Source Software (2005): n.pag.

• Lerner, Josh and Jean Tirole. “The Simple Economics of Open Source.” Journal of Industrial Economics 52 (2002): 197-234.

• McCarthy, Jim. Dynamics of Software Development. Redmond: Microsoft Press, 1995.

• Montesquieu, Charles Louis de Socondat. Persian Letters. Bainbridge Island: Bainbridge Island School District, 1721.

• Raymond, Eric S. The Cathedral and the Bazaar. Sebastopol: O’Reilly, 1999.

Works Referenced

• DeMarco, Tom and Timothy Lister. Peopleware. New York: Dorset House, 1987.

• Lum, Rosalyn, Laurie O’Connell and Alexandra Weber Morales. “The Open Road: A History of Free Software.” Software Development Jun 2005: insert.

• “Meritocracy.” Wikipedia. 15 May 2005 http://en.wikipedia.org/wiki/Meritocracy.

• Raymond, Eric S. The Cathedral and the Bazaar. Sebastopol: O’Reilly, 1999.