The book is a collection of essays written by Eric S. Raymond advocating for open source software development. I’ll focus this review on my favourite essay: The Magic Cauldron. The essays should be readily available online in case you are interested.
For context, in a prior essay whose title shares the name of the book, The Cathedral and the Bazaar, ESR introduces 2 models of software development: the cathedral model (closed-source) and the bazaar model (open-source). He then argues in that essay why the bazaar model is superior because open development and decentralized peer review lower costs and improve software quality, using Linux and his own open source experience as evidence. Assuming this is true, the obvious other problem with open-source is that it does not seem like a viable business model. It is this exact belief that ESR challenges in The Magic Cauldron.
At the time of writing (1999), the software industry operated with the same business model as a factory - producing software and selling it at a markup. ESR points out numerous problems with this model. Firstly, the value of software depends heavily on future vendor support and maintenance. Secondly, oftentimes the software is not as important as the underlying content. For example, suppose you have a real-time stock market data service (think Bloomberg API). Your server software is not as valuable as your proprietary data feed. By open sourcing the server software, you can lower development costs, improve code quality, and more easily provide long-term vendor support. Now all that’s left is to charge a subscription for access to your data feed.
A more recent example that comes to mind is Anthropic’s Model Context Protocol (MCP). Open sourcing MCP is a really interesting decision. Keeping it proprietary certainly would have given Anthropic a huge edge over its competitors for anyone wanting to build an AI agent. Instead, by open sourcing it, Anthropic grew the entire market for agents, forced their competitors to adopt their standard, and cemented Claude as the natural choice of foundation model to use.
TL;DR This guy is a real critical thinker, a great essay writer, and basically predicted SaaS in 1999.