Variation and Analog Computers

What We Can Learn from the Norden Bombsight

Editor’s Note: I’ve been delayed in writing at my usual cadence owing to taking on a new role helping a startup implement whole-system improvements based on Dr. Deming’s teachings and some that he inspired. Thanks for your patience as I work through my backlog of posts to finish up and get published. Also, if you have ideas for topics you’d like to see explored, please let me know in the comments. - CRC.

FOR TODAY’S POST, a surprise that I wasn’t anticipating writing, but was inspired to by the above Veritasium video presented by host (and fellow Canadian) Dr. Derek Muller. Incidentally, if you’re at all curious about math and physics or have a budding scientist in your family, Muller’s approachable deep-dives are well-worth binge-watching: I always come away from his videos having learned at least one or two new things.

However, I digress: What piqued my interest to share this particular video was Dr. Muller’s retelling within of the US military’s investment in the failed Norden Bombsight, which immediately brought some connections to Dr. Deming’s philosophy to mind - from the video:

Designed by the eccentric Dutch engineer Carl Norden, the Norden Bombsight was built to enable high-precision airborne bombing. It implemented 64 different simultaneous algorithms, including one that compensated for the rotation of the Earth as the bomb fell.

The Norden was one of the most closely guarded secrets of the war. To prevent the technology from falling into enemy hands, American bombardiers carried handguns specifically to destroy it in the event of a crash.

But despite its hype and funding, the Norden didn't work as advertised. With over 2,000 fine parts, it required extreme precision to manufacture. The problem with analog computers is that the physical device is a model for the real world. So any inaccuracy in the components translates into inaccuracy of the computation. And since there will always be some slop in the connections between parts, if you run the same calculation twice, you won't get the exact same answer.

In the American campaign against Japan, bomber crews using the Bombsight were unable to destroy critical Japanese war infrastructure, and ultimately the US abandoned its precision bombing approach and instead blanketed whole Japanese cities in napalm.

Veritasium - The Most Powerful Computers You Never Heard Of. December 21, 2021. (14:45)

Here’s some take-aways I had from this when considered through the lens of Dr. Deming’s philosophy on management, as summed up in his System of Profound Knowledge, along with some historical observations:

• First, there’s some irony: During the war, Deming consulted to the military on quality improvement through statistical control, which he taught extensively. Even at this early stage in the development of his thinking some fifteen years before he’d hear Dr. Taguchi deliver his paper in Japan, he would have been able to diagnose problems with the Norden Bombsight as originating in the system of manufacture as a whole, as evidenced in the ranging specification tolerances of the parts.

• Second, and in reference to a lesson Dr. Taguchi would impart through his theory on loss functions, quality is not defined by parts taken separately that “meet specifications” but are centred on the quality “target” with minimal variance. (This was discussed at length in my October 25th newsletter on The Taguchi Loss Function). This requires “aiming” the corresponding processes by first understanding their averages, then working to manage the corresponding variation toward the “target”. This requires thinking about the interactions or gaps in the processes that contribute to the gaps in the parts.

• Third, and in relation to the prior take-away, the failure to “mind the gaps” in how the parts of the Norden Bombsight worked together as a system led to multiple failures to keep centred on the quality “target”. Dr. Taguchi would refer to these misses on the curve as not only losses for the product and customers, but to society as a whole. In this instance, the unreliability of the bombsight would lead directly to the abandonment of precision targeting of factories and the development of the atom bombs dropped over Nagasaki and Hiroshima: A catastrophic loss to society by any definition.

• Fourth, I was given some pause to think that were it not for Carl Norden’s failures with his bombsight, there may never have been a need for Homer Sarasohn, Dr. Joseph Juran, or Dr. Deming to travel to Japan to help rebuild their manufacturing capacity, and possibly no need for Dr. Taguchi to deliver his Tokyo paper in 1960, or for Deming to be there to hear it. Consequently, all that we know now would have never been. At least, not as we understand it today.

Reflection Questions

Consider the beleaguered Norden Bombsight with its ambitious complexity and aim: It was designed with the intent to minimize casualties and enable precise targeting of factories and infrastructure, yet failed and literally caused the nuclear option to be exercised as a consequence of having a design that outstripped the capacity and capability of the systems that were required to manufacture it. What analogs have you observed in your organization where overly-ambitious efforts with many moving parts designed to solve a problem were undertaken with similar results? What happened next? Would some knowledge of thinking about the interactions between the parts have averted disaster, or was there too much organizational inertia to overcome?