The Cathedral and the Bazaar

The Cathedral and the Bazaar is an essay by Eric S. Raymond that gave the free software world a model for understanding its own way of working. First presented at the Linux Kongress in May 1997 and later expanded and published online at Raymond’s site catb.org, it argues from direct experience that a particular style of open, distributed development consistently outperforms the conventional closed approach.

The essay’s central metaphor contrasts two models. In the cathedral model, source code is developed by a tightly controlled group of developers and released only when finished, much as a cathedral is built to a master plan behind scaffolding. In the bazaar model, exemplified by Linux and Linus Torvalds, development happens in the open with frequent releases, a large pool of co-developers, and constant public feedback, resembling the noisy, parallel activity of a great bazaar.

Raymond did not merely observe the bazaar from outside; he tested it. The essay narrates his own experiment running an open project, the fetchmail mail-retrieval utility, deliberately adopting bazaar practices to see whether the model would work in his hands. From that experience he distilled a series of lessons, the most quoted being that “given enough eyeballs, all bugs are shallow,” meaning that with a large enough body of testers and co-developers nearly every problem will be diagnosed quickly by someone.

The timing of the essay proved historically important. Released frequently in revised form and widely read, it circulated just as Netscape was weighing what to do with its losing browser business. Netscape executives have credited The Cathedral and the Bazaar as a factor in the company’s January 1998 decision to release the source code of its browser, the move that launched Mozilla and helped trigger the meeting where “open source” was coined.

More than a technical paper, the essay was a persuasive argument that openness was not charity but engineering advantage. By framing distributed development as a superior method rather than a moral stance, it helped open the door to the pragmatic, business-oriented branding of “open source” that would follow within a year.