ISBN 0801845882
Buy it from Amazon.com
Buy it from Amazon.co.uk
What is engineering knowledge anyway? Is engineering just Science's poor cousin, Progress wrapped in an oily rag, or is there something distinctively different about it?
One answer, of course, is that engineering must be practical. Theories can be fine and elegant, but if the plane flies and the bridge stands up, the engineer was right regardless of how the result was arrived at.
But Vincenti goes about his task in a more satisfying way. An aeronautical engineer, he's taken a careful retrospective look at five aspects of the growth of knowledge in aeronautical engineering between 1908 and 1953 -- the time when things were uncertain, and people were feeling their way sometimes almost entirely in the dark towards safe, reliable, predictable ways of making aircraft.
Vincenti warns that the stories will not be easy going, and this may be true for historians coming to his material. But engineers will probably not find the technicalities at all difficult; the story depends on them, of course, and the tale is fascinating, gripping as a good novel. It's all the better for being both true and widely applicable to the questions that arise in development work such as requirements definition today. In fact, one of the 'stories' is precisely about 'The Establishment of Design Requirements'. Not surprisingly, it's the longest and most complex of the chapters, but it was almost impossible to put down. It is rich in details, often astonishing, sometimes funny:
In an exchange of penciled intraoffice memoranda, Captain Hatcher let down his hair with a candor that rarely appears in more official documents:At present we simply specify that the airplane shall be perfect in all respects and leave it up to the contractor to guess what we really want in terms of degree of stability, controllability, maneuverability, control forces, etc. He does the best he can and then starts building new tails, ailerons, etc. until we say we are satisfied. Unsatisfactory airplane characteristics were doubtless due in part to the inability of designers to design for what was wanted. As indicated by Hatcher's memo, however, what was wanted was far from clear. (p77)
Couldn't happen today? Well, what is a 'fitness for purpose' clause, then?
Vincenti is masterly in showing how engineers fumbled towards knowledge; few writers have attempted anything like this. Sometimes, Vincenti shows, the community was led astray by prejudices held by special groups of stakeholders (as we'd now say), especially pilots; sometimes by common sense; sometimes indeed by the apparent implications of (correct) mathematical modelling!
Most startlingly, perhaps, it turns out that stable, reusable requirements are the OUTPUT of years of effort trying to understand the properties of different designs in a domain. The requirements epitomise the laws governing a wide range of possible designs (not all imaginable aircraft: helicopters for instance certainly behave quite differently). What price 'solution-independent' user requirements, then? Today's aeronautical engineers see it as obvious that manoeuvrability depends on stick-force-per-g, ie that a manoeuvrable fighter must respond to a light pull on the stick. But that wasn't at all obvious for 25 years of research and development! Perhaps there are things we know but can't explain – along the lines of Polanyi (in fact, Vincenti does mention that skilled tasks such as riveting 'can't be learned from the book' but are in the 'neuromuscular skill' of the practitioners) and important practical things that were known that are being forgotten. Reflection on practice – as Schön advocates – is valuable, but easier said than done.
This is a wonderful book, written with an engineer's precision, a historian's eye for detail, and a storyteller's clarity. It is impossible to read it without being struck by one reflection and insight after another. Learning is difficult, when you're doing it for the first time! And of course, every project is a new set of discoveries. Students and experienced engineers will find Vincenti invaluable. It is a book to muse over, and to come back to.
If you are still not sure you must read Vincenti closely, take note of what Michael Jackson said about it, when concluding his keynote lecture to the world's requirements engineers in Barcelona:
"Read Vincenti's book. Read it carefully. Read it one hundred times."
© Ian Alexander 2005, 2008
You may also like:
