Sunday, May 8, 2011

The Design of Design by Fred Brooks

If you haven't read Fred Brooks' The Mythical Man-Month about software engineering and project management (updated in 1995) you should.  While I enjoyed his 2010 effort The Design of Design, it didn't have the impact of its predecessor.

It's not that The Design of Design is a bad book.  It's a collection of essays about designing things (whether it be software or a house) and how design can be taught, learned, and fostered.  But the nature of the essays gave the work a fragmented feel.  You'd think having a book of digestible chunks would make it easier to read, but after a chapter here and a chapter there over a period of weeks it took some dedicated time on an airliner for me to finish it.  Plus, I skipped the last of the book's six parts consisting of seven case studies of various designs.  Frankly, I wasn't motivated to learn about how Brooks added a room to his house.

But the first five parts are worth reading.  And I'm not saying that just because he mentions CFD (computational fluid dynamics) by name as an example of how advanced the designer's toolkit has become.  Brooks wrote something that really struck a chord with me: "My whole intellectual life has been one of throwing passionate subfield interests overboard as they have exploded beyond my ability to follow them."  I've thrown enough overboard that my id can cross the Pacific Ocean of my ignorance without getting its feet wet, skipping easily from one bit of mental flotsam to the next.

Part I of The Design of Design lays groundwork for the remainder of the book by reviewing the basic nature of designing and models for its implementation.  Brooks draws a distinction here between design and the design process for the simple, practical fact that few of us get to do truly unique designs (think Apollo) and instead are designing new versions of things that already exist.  Therefore, the design process is a valuable tool that facilitates our ability to accomplish real work.  Brooks describes different process models including rational, waterfall, bazaar (as in The Cathedral and the Bazaar) and spiral and is very clear with his opinions:  "I find Boehm's Spiral Model the most promising."  A chief takeaway for me was something I've seen validated over and over again in the software business: a designer's goal is to help clients decide what they want designed.  Stated differently, one should not assume that requirements and constraints are known in advance.  The design process often illuminates them.

Collaboration and telecollaboration are addressed by Brooks in Part II.  He points out that while the move toward team-based design is well-founded due to the complexity of the modern-day design task, all great designs throughout history succeeded because of conceptual integrity that comes from having one chief designer.  That is the main challenge to be faced by design teams.  Drawing a bit of a parallel with Patrick Lencioni's Death by Meeting, Brooks quotes Baruch who wrote "A meeting is a refuge from 'the dreariness of labor and the loneliness of thought'."  In essence, meetings are an essential element of the design process, not to be feared or even considered a necessary evil.

The chapters in part III tackle design perspectives, specifically the rational versus empirical approach to design.  Rationalism, the ability to design correctly based on thought alone, is rejected due to the simple fallibility of people.  Once empiricism is embraced, Books leads us through discussions of user models ("An articulated guess beats an unspoken assumption."), budgets (name, track, control and publish the critical resource), constraints (form is liberating), aesthetics (consistency, clarity, and style), and exemplars.  He delves a bit into the dark sides of design with how designs go wrong, how designs get divorced from reality, and then into tools for tracking a design's trajectory.

Part IV is where I started to lose interest.   I just couldn't keep up with "A Computer Scientist's Dream for Designing Houses."

Part V is a nice bookend to Part I in that Brooks reminds us that great designs come from great designers, not great processes.  While the latter is necessary to the success of the former, it is incumbent upon us to ensure that we foster the development of design skills from college education through professional development.  This echoes thoughts from an article in Mechanical Engineering magazine from a couple month's back about how a student's grades in design courses should be used as a measure of their success in the profession.

Overall a decent read.  If The Mythical Man-Month warrants a rating of 9 then The Design of Design is about a 6.5. But Brooks' thoughts are worth reading if only to help clarify your own ideas.

I received no compensation of any kind for this review.


Jim said...

Thanks for the review - I just put this on my "hold list" at

Did Brooks mentioned any Buckminster Fuller designs? In particular, when I was in Detroit for a partner event, I budgeted a free afternoon to see the Ford museum in Dearborn. One of the unexpected things I saw was the Dymaxion House by Buckminster Fuller. Many aspects were brilliant (bathroom + water use), though there were some practical flaws (lack of insulation).

John said...

Brooks did not comment on Fuller to the best of my recollection.

Francis Shivone said...

Good review. It make not be always true but the statement, "A meeting is a refuge from the dreariness of labor and the loneliness of thought," is still good enough to quote.

BTW -- I went down to the Kimbell to see the big hole and walk a round a bit. I had never known that the current main entrance was not designed to be the main entrance. Worth a trip if for building enthusiasts.

John said...

Lencioni's Death by Meeting ( makes it clear that for most of us, meetings are our work product - where decisions get made and therefore where it is important to have well run meetings.

I haven't been in The Kimbell in years. My limited art time is consumed by The Modern.