Open source and the mythical man month
There's very little code here. Just straight talk on managing software projects, from an old hand.
Many of Brooks' examples from from the System/360 project in the 1960s, which he managed for IBM. Some are older.
Much of the book is a lesson in how completely software development has changed since you were a kid. Even if you're still in your 20s.
The book is most famous for Brooks Law, the idea that adding people to a project just makes it run slower, as communication costs add up.
What got the attention of my programmer-wife were two chapters he added to later editions, titled No Silver Bullet and No Silver Bullet Refined. In it Brooks argues for what I call Moore's Law of Software, namely that there is no Moore's Law of Software.
Software is as much art as science, and remains a hand-made product, despite all the many advances Brooks names and discusses with such simplicity and grace.
In the two Silver Bullet chapters Brooks looks at many contemporary subjects, at object-oriented programming, at artificial intelligence, even at the impact of the mass market.
What he does not discuss is open source.
He does touch on some trends open source makes possible, such as incremental development and code re-use. But even in this decade most breakthroughs remain the work of individuals or small teams -- think iPhone and Twitter. Big projects can be overwhelming.
It's this that makes open source, and the open source process, so powerful. Open source does not repeal Moore's Law of Software, but it enables a lot of re-use, it provides a management structure for really vast projects, and it creates a virtual market of code that anyone can participate in.
I need to note here that shrink-wrapped products and the creation of software mass markets are important productivity drivrers in Brooks' book. Both can be delivered through a completely proprietary model, as we've seen for instance at the Apple App Store.
But it's the Internet, and the open source business models which emerged from the Internet, that are the closest thing to a revolution we have ever had in software development. Brooks reminds me just how new that revolution is and just how further it has yet to go.
The Mythical Man Month reads like history, but like any great book its lessons speak to our time and beyond. If you have a student graduating this month who hasn't read Brooks' book, it makes a great present.