The Cathedral and the Bazaar

The title essay began as a talk in 1997 and spread across the internet before it was ever a book. Eric Raymond’s argument was simple to state and hard to ignore: software built in the open by many contributors can beat software built behind closed doors by a plan.

What it’s about

Raymond contrasts two ways to build software. In the cathedral model, a small group works in isolation and releases a polished version at intervals. In the bazaar model, development happens in public, releases come early and often, and a crowd of users and contributors find and fix problems. He uses the growth of Linux under Linus Torvalds as the main example of the bazaar, and describes his own experience running an open-source project, fetchmail, as a case study.

The essay popularized a set of rules of thumb, including “release early, release often” and the line he named Linus’s law: given enough eyeballs, all bugs are shallow. The book collects the title essay with related pieces on the culture and economics of open source.

The timing mattered. The essay was widely read around the time Netscape announced, in 1998, that it would release the source code of its browser, and it is often credited with helping persuade companies to take open source seriously.

Why it belongs on coldboot

It is the best-known attempt to explain what happened with Linux and why it changed the industry. It also fits the story told in the GNU Emacs and GCC articles, which are earlier chapters of the same movement.

Keep in mind

This is an argument, not a neutral history. Raymond’s description of how Linux was developed is a generalization that later historians have complicated, and his examples of the “cathedral” model are debated. It is most valuable as a record of how people at the time understood what was happening.

Edition linked: revised edition (O’Reilly, 2001).