Book review: The Cathedral & the Bazaar

A review of The Cathedral & the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary, by Eric S. Raymond. Published by O'Reilly & Associates, Inc., 1999. Foreword by Bob Young. Hardcover, 288 pages. ISBN 1-56592-724-9. I reviewed the first edition, October 1999.

Published in the Dallas/Fort Worth Unix Users Group Newsletter, April 2000

Note - I use the term "hacker" to describe the people who participate in Open Source projects, using the same definition that the book's author uses. If instead you use the common media interpretation that hackers are snot-nosed kids who like to break in to secure servers, you're going to miss the point.

This book is a collection of several of Eric's papers about the Open Source model, along with some other introductory and supplemental material. The papers included are:

"A Brief History of Hackerdom"
"The Cathedral and the Bazaar"
"Homesteading the Noosphere"
"The Magic Cauldron"
"The Revenge of the Hackers"
"How to Become a Hacker"

The papers in the book are well-written and integrated fairly well together. A few concepts seemed to get jumbled a bit as I worked my way toward the back of the book, but given that the papers were all written at different times, the book flows pretty well. The flagship "Cathedral and the Bazaar" paper is well polished, though there are a few scattered grammatical flubs and a few misspellings in the newer chapters.

In "A Brief History of Hackerdom," Eric gives a concise history of the hacker culture, which is all tied up in the history of Unix and the Internet, though the origins predate both. Longer histories are found in Steven Levy's legendary book Hackers: Heroes of the Computer Revolution, and the less-known book A Quarter Century of UNIX, by Peter Salus. In this chapter, Eric also mentions the Jargon File, which was one of my first introductions to the hacker culture, and is required reading for any self-respecting hacker (I also reviewed the book based on this). Hackers seem to have a lot of pride in their history. Just ask one about the "good old days" of plug boards, punch cards, 2400 baud modems, or 486 processors, depending on when they joined the ranks of hackers, and then find a comfortable chair because you're bound to start a protracted session of reminiscing.

The next chapter is "The Cathedral and the Bazaar," after which the book is named, and the first of three papers in a series. In this chapter, Eric explores the Open Source phenomenon that has grown out of the hacker culture, mostly as it relates to the development of Linux. Well, more accurately, he discusses the bazaar-style development model, of which Open Source is one example. He describes a case study where he took over the development of what came to be known as "fetchmail," and gives some very useful insights into the mechanics of a bazaar-style project.

One thing that struck me from the "Cathedral and the Bazaar" chapter was a concept that Eric related from a friend of his, that the hacker culture only accepts the most talented 5% of the population of programmers. It's Eric's hope that use of the Open Source model will have a tremendous amount of growth as the Open Source concept goes mainstream. But if the labor pool is limited to 5% of the possible programmers, then Open Source can only grow as fast as the total number of programmers grows. What kind of growing pains does Open Source have ahead if projects need to draw from, say, the top 20% in order to grow as fast as everyone wants it to? I guess it remains to be seen whether the Open Source movement is scalable, without resorting to using some of the management techniques that Eric bashes as unnecessary overhead. (Eric tells me that he's not sure whether the 5% theory is correct or not.)

Next up in the book is the eclectically named "Homesteading the Noosphere," the second in the trilogy. Here Eric continues to explore the hacker culture, relating the customs of Open Source software ownership to the practice of homesteading. He also discusses how hacker culture is a "gift culture," which turns out to be a very good model for explaining why things work the way they do on Open Source projects. A couple of things in this chapter could use some additional explanation. For example, the potlatch and multi-millionaire reference on page 99 could use more elaboration. And on page 112, I'd like to see further discussion of why hacker leaders exhibit self-deprecating behavior in order to avoid a personality cult.

Closing out the trilogy is "The Magic Cauldron," where the author talks economics, including several funding models for open source development. It's a good expansion on the case study in Open Sources, "How Red Hat Software Stumbled Across a New Economic Model and Helped Improve an Industry." This is the chapter that might convince you to buy the book so you can show it to your CFO, because having it in print will be more convincing that pointing your management to some random URL. Eric delves into some serious economic theory here, though making a complete business case based on these concepts will still require a good deal more work. The author does a good job of dispelling some myths, and he doesn't give special preference to ideas that originated in the hacker community. His critical thinking about both the cathedral model and some commonly held beliefs about the bazaar model lends credence to his arguments.

Capping off the book is "The Revenge of the Hackers," where Eric describes the hacker culture's rise from obscurity to mainstream attention, and he chronicles his own role in the marketeering of this emergence. He includes a good deal of Microsoft-bashing, including a self-assured prediction that Windows 2000 would be either canceled or totally unusable when it ships. There's a passing reference to "memetic engineering" on page 200 - if you're not familiar with memes and Richard Dawkins, you'll probably miss it. It has been several years since I've read any details about memes, so I was surprised to find a considerable amount of information on the web now about memetic engineering. It's an interesting diversion. Anyway, back to the task at hand.

In this paper and throughout the book, Eric mentions the release of the Netscape source as Open Source (the Mozilla project) as a crucial turning point for Open Source becoming mainstream. Unfortunately, two years later, Mozilla still doesn't have a beta-quality browser. I wrote this article and did some of my research using Mozilla, but doing so was difficult because of the number of bugs I ran into. The good news is that other companies have announced that they would support their product on Linux or would release their code under an Open Source license. So maybe the revolution isn't so dependent on Mozilla's success now. I don't have a good update since the paper was written, so I'm curious about where the Open Source revolution stands at the moment.

Don't stop reading when you get to the appendices. "How to Become a Hacker" is a real gem. After reading this piece, I realize that I picked up a lot of the concepts when I became immersed in the culture myself. You aren't a hacker until others recognize you as a hacker. You don't seek hackerdom; it seeks you. Except when I first got started, we also made note of the ranks further up the scale, using terms like "guru" and "wizard." Anyone who described themselves with these terms was discarded as a phony. For anyone making a conscious decision to join the hacker culture, or for any hacker who wants a better understanding of the culture they're a part of, this paper gives specific suggestions for how to become a valuable member of the hacker community, all the way down to suggesting which programming languages to learn first..

The book uses fairly wide spacing between the lines - probably a trick to make the book thicker and appear more substantial. However, this is preferable to another trick often used by publishers, which is to make the author pad the book by writing worthless fluff. There's no fluff here, and the price is on the lower end of the scale, so having it a little bit thin on content versus similar books doesn't bug me. I think I'm spoiled by O'Reilly lay-flat soft cover binding. The hardcover binding on this book is stiff and it's difficult to keep the book open.

In conclusion, I think the papers in this book are required reading for anyone who wants to work within or understand the Open Source movement. Whether you buy the book, or make do with the 95% of the content that's available online is up to you. But you're going to have to demonstrate your hacking skills by finding the online content yourself. :-) Having the ideas available in book form will serve to legitimize them to someone who doesn't yet grok the whole thing. And that's the whole point of the exercise - making sure more people "get it."