One of the less well-known software engineering classics, “Winning with Software“, seems to have lost a bit to the current agile mainstream. But even for successful agile adoption, there is a lot to learn from this apparently waterfall-heavy masterpiece.
I have to admit I didn’t study the book lately in the detail it deserves. That’s because I was lucky to meet its author, Watts Humphrey, in person three years before he died and had the unusual opportunity to get the ideas straight from the source. ~ fifty years of industry experience… That was impressive. He said this book reflects the essence of our conversations.
Anyways… the book has the subtitle “An Executive Strategy”, and that is well deserved: it belongs to a series of three and focusses on a project sponsor’s perspective on running software development projects: rational management (i.e. relying on data), the importance of quality, leadership, motivation, teamwork… It could be any management book, except that Watts was focussing on what made software different from hardware: the fact that software – unlike hardware – doesn’t know of real production preparation and production steps. As a consequence, many of the standard quality management practices in engineering can’t be applied (because they are localized in production preparation and applied to production – if these two are gone, there’s nothing to attach them to). So he had to figure out new ways of managing quality and he did. Iin this book (along with its two companions) he explains how to manage quality even without production preparation and production.
His emphasis and his approach of making software development measurable is with me every day, even though I usually rather rely on agile practices to get the job done. But how we measure progress is certainly interchangeable so long as we understand how to use the measured results: to accept the measurements whether we like the outcome or not, and to derive sober consequences from them. And that is what this book is largely about. It starts with laying down timeless principles of software management and illustrates them beautifully with examples from the research of the Software Engineering Institute (SEI):
- Recognize that you are in the software business
(so treat it as such) - Quality must be your top priority
(SEI research shows that quality issues are overwhelming everything else, and I’d like to add: the SEI includes all artifacts and communication on the path from idea to product in this consideration) - Quality software is developed by disciplined and motivated people
(and I’d like to add that scrum or other agile methods, if done right, are very disciplined. That the discipline itself differs from PSP/TSP is largely irrelevant. There are many ways to skin a cat.)
On top of that, many of today’s “modern” management wisdoms such as the essence of Daniel Pink’s “Drive” are omnipresent in Watts Humphrey’s work, such as:
For any but the simplest work, fear and greed are not effective motivators.
Daniel Pink’s entire book “Drive” and his famous TED talk build on the idea incorporated in this one sentence, and there is no end to examples like this.
Oh, and by the way: so far, we have spoken just about the first half of the book. The appendix (the second half of the book) is dedicated to the implementation of Personal Software Process / Team Software Process (PSP/TSP), the Software Engineering Institute’s then-approach to making software development predictable. This is the aspect that really seems a bit outdated in our agile era. And even so, if I had to realize, say, the control logic for a nuclear power plant, I’d still use PSP/TSP without a moment of hesitation if there is a suitable team.
As with “The Mythical Man-Month“, many of the concrete tools don’t apply anymore. After all, this book is almost 20 years old. Nevertheless, the intentions and philosophy behind the book still hold. It’s like the famous “a fool with a tool is still a fool” backwards: wisdom is still wisdom, even when the tools change. And the tools will change, at least for another decade or two, but Watts Humphrey’s practical wisdom remains.