*Empirical Structural Design for Architects, Engineers, and Builders* by Thomas Boothby is exactly what its long title suggests: a textbook on how to design without modern analysis. Empirical design can mean a lot of things, but on page 1, Mr. Boothby makes it clear what he wants to talk about. He gives an example for wood joists that shows that simple depth-to-span ratios can substitute for stress analysis. There are, of course, caveats, but this example and the many that follow in a discussion of nearly every type of common building structure make it clear that there are other ways of approaching structural design than complex calculations. Here’s my caveat: the empirical designs here are all great as a first-pass at design. The final design must be based on the standards of modern-day structural design, which is to say, have to follow our design codes. A rational way to use the empirical designs is to establish a goal which can then be designed per code. This is faster than using iterations of code-based design and more precise than using only empirical design. A good example would be to say that a simply-supported steel floor beam supporting a concrete slab should have a depth approximately 1/24 of its span, and so start a design for a beam with a 20 foot span by trying a ten inch deep (W10) shape. This is in table 5.5 in the book and was taught to me at my first job, in 1987, as “For your first try, take the beam span in feet, divide by two, and change the units to inches.”

I highly recommend this book for engineers looking for a shortcut to a first approximate design that can then be analyzed more precisely, and to others (the architects and builders of the title) looking for a way to approximate a structural design. But there’s more…

This is probably a good time to mention I know Tom Boothby a bit, and he’s a friendly and inoffensive man, which is why it’s darkly amusing to see him deliberately trigger a new battle in the 200-year-long war among engineers over the merits of empiricism and analysis. All engineering was empirical if you go back far enough, since there was no theoretical basis for what is now called “rational analysis” of forces and movements. Various geniuses – Galileo and Euler immediately spring to mind – worked on establishing mathematical ways to explain structural behavior and gradually built up groups of equations and methods that could be used with reasonable accuracy to describe structural performance. The existence of analytic methods didn’t mean they were used, but in the early 1800s the analytical camp took shape in Europe. Arguably the leaders of that camp were the French schools – the École Nationale des Ponts et Chaussées and th*e *École Polytechnique *– *and their graduates. For a lot of the 1800s, the analytical school was concentrated in continental Europe while the empirical school was concentrated in the UK and US. There were many individual engineers who did not follow the national pattern, of course. The development of analytical tools continued, and by 1950 the analytical camp had more or less won. The development of computers that could perform endless calculations moved things further towards analysis by enabling mathematic tools (the finite-element method is the most obvious) that were simply unrealistic to perform by hand.

And yet, the discussion continues. Many prominent engineers have written some version of a lament about “those kids today” relying on computers rather than hand calcs. While I understand that sentiment, it seems to me to be a clear outgrowth of teaching analysis at the expense of empiricism: if everything is a mathematical analysis, why shouldn’t you use the computer, which is going to perform the math faster and more accurately than you will? A discussion of empirical issues is far more useful to young engineers’ educations and old engineers’ error rates than forcing people to work calculations by hand. In a new building, what are the load paths? Where is the stiffness? Which connections will be highly stressed because they are between stiff and flexible elements? In an investigation, which cracks are advancing and which are working back and forth? If you think there’s structural movement, can you trace it though all parts of the structure through cracks or deformations?

So this book is important not just for what it contains (a useful guide to empirical design for people who have not seen it before) but for what it represents (a school of thought that once dominated engineering and is now a niche, but which is still meaningful and useful). That may be a lot of weight of expectations to put on its back, but I suspect it can take the load.