The Design of Design - Frederick Brooks

Frederick P. Brooks, Jr.

March, 2010
ISBN: 978-0201362985

I had traveled to Dublin, Ireland where Fred Brooks - project manager of System/360 and author of The Mythical Man Month - would be a featured speaker.

System/360 made IBM the 1,000 pound gorilla of computers in the age of mainframes. People today don't remember that back then, "computer" and "IBM" were almost the same thing in most people's minds. In my book, Fred Brooks deserves the credit for making that happen. I have written, "Fred Brooks is my personal hero. If I could make the decision, there would be a monument to him in every park in America."

I wanted to tell Fred about the software development project that I was now managing. It was the largest development project in the company. In a very real sense, the jobs of several hundred people were on the line with the project. But over a million dollars of the budget had already been spent with little to show for it and the project was spiraling down the drain. Top management wanted to outsource the whole I/T department. I/T management proposed a comprehensive reorganization instead. Part of the deal was that I was put in charge of the project.

This project was consuming my life like a fire consumes dry wood. I was reaching out almost as a drowning man. Fred agreed to talk to me about it for a while. I told him about the project and the nightmarish problems I was facing. He listened carefully, and at the end told me, "I've seen this before. There's not much chance that the project will ever succeed."

That wasn't quite what I wanted to hear just then.

Fred was more inspirational than he knew. I attended every session he was in at the conference. In a world of international academic politeness, Fred was direct to the point of bluntness in everything he did. When he chaired a panel discussion, he announced that questions and comments from the audience were encouraged, but everyone should state their name and affiliation first. The first three or four questions followed his rule perfectly. But then an excited man jumped up to add his opinion to a previous point. ... Fred cut him off like the voice of God. No one forgot again.

The Mythical Man-Month added "Brooks' Law" to the culture of software development, "Adding manpower to a late project makes it later." When management suggested that adding more people would help my team, Fred's inspiration helped me tell them no.

He wrote No Silver Bullet for that conference. (I still have the twenty pound book of papers presented at the conference.) Fred wrote that the core of software development is "essential complexity" and most of what software engineers do is focused on this unavoidable complexity. Software development is hard and always will be. For my project, management had already bought into a high-priced "Computer Aided Software Engineering" system before they put me into the project. Fred helped me tell them it wasn't going to make that much of a difference.

In The Design of Design, Fred continues his long tradition of unapologetically putting his vision of software design in plain view, even though it might not be what we want to see. The people paying the bill for a software project passionately want to believe that when half of the money has been spent, then the project is half finished. Fred helps us understand why that's usually not true.

And while it might be uncomfortable for those of us who have software development skills somewhere in that vast middle ground, Brooks points out that while there is "No Silver Bullet" in software development technology, there are silver bullet people. Here's what Fred says about it:

"The boldest design decisions, whoever made them, have accounted for a high fraction of the goodness of the outcome."

"Most of the great works of the human mind have been made by one mind, or two working closely."

At the conference, I remember one person pointing out that studies had shown that the lowest and the highest productivity people in an organization will ordinarily (That's "most of the time", folks!) be different by a factor of ten or more. The person asked Fred, "What do you do about those people on the low end?" Fred said that the only answer was to get them out of your project.

Software development is full of people who sound great, but have no reality behind them. Fred is all about the reality. The main reason for reading this book is that it's the most thoroughly grounded and realistic thinking you're going to find in print.

Fred was right ... and wrong ... in his prediction about my project. Three very, very long years after he said it would fail, my team put the project into production and it did everything it was supposed to do. But in a more real sense, it did fail. In the time it took to finish it, the sponsors who originally wanted the project had retired and the part of the company that would use it was reorganized. Like a Frankenstein monster, the system staggered around for a while and was then destroyed.