“The desire for safety stands against every great and noble enterprise.” ― Tacitus

I remember back in the days of my undergraduate education leaning about safety factors for the first time (sometimes the “fudge” factor, where “fudge” is a polite way to say another word!). It was just an accepted practice to account for what isn’t completely known and is a generally acknowledged good idea. The basic idea is to pad images copyout the margin of safety on various design decisions or analyses to account for the possibility that the formulas being used are off, or maybe you made a mistake, or forgot to account for something. It’s all the unknowns, including the “unknown unknowns” we keep hearing about. The true depth of the concept is simply passed over by the undergraduate instruction where a bit more contemplation would serve everyone so much better. It would serve to inject some necessary humility into the dialog and provide context for vibrant research in the future.

“Reports that say there’s — that something hasn’t happened are always interesting to me, because as we know, there are known knowns; there are things that we know that we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns, the ones we don’t know we don’t know.” – Donald Rumsfeld

images-1When I was first introducted to safety factors it wasn’t treated like anything profound or terribly important. Just a “best practice” with no philosophical context for what it means. Upon my return to this topic in the context of recent research, it began to dawn on me that safety factors are really deep and conceptually rich. The fact that there is simply a lot that we don’t know, or even know we don’t know is potentially terrifying. We could be screwing up horribly and not even know it. At some other level what the safety factor is really telling you that the vaunted research embedded in the formulas you’re applying is potentially wrong or inapplicable to your situation. What the professors didn’t say to you so plainly is that the research of the greats might be just plain wrong. In other words, the greatest research in their field might not be worth completely trusting.

This is really OK. We are all adults here, right?

“Risk means ‘shit happens’ or ‘good luck” ― Toba Beta

The whole concept of safety factors is a way of pointing out where work is needed. A too generous safety factor is wasteful and expensive. Of course, a key tension is that a safety factor that is too small is dangerous. This is where accidents and disasters with technology lie. It is worth pointing out that to some degree accidents and disasters are unavoidable, and sometimes the events in the tail of probability distribution happen. No safety factor is going to prevent this. Some misery is unavoidable. Sometimes the problem is deeper than what a safety factor can deal with. The universe is complex and nonlinear with lots of abnormal things. Most of our understanding of the universe is linear, simple and normal (statistically). Our best science is setting us up for a fall. Not often, but often enough that we shouldn’t be too surprised when it happens.

“On a long enough timeline, the survival rate for everyone drops to zero.”- The Narrator from Fight Club

This isn’t in any sense an attempt to belittle the accomplishments of the generations of scientists whose work set the foundation for the modern World and its wonders. Humanity has accomplished so much, and the understanding of the World we have today is an immense virtue, but let’s not get carried away. We are not masters of the Universe, it is the master of us.

“We have to remember that what we observe is not nature herself, but nature exposed to our method of questioning.” – Werner Heisenberg

My point is to throw down the gauntlet at those who think science is done, and we understand everytUnknown copyhing. The truth is we don’t know shit. So many important things are simply mysteries, and as we unveil their secrets, new mysteries will rise to take their place. We never really know the truth, we just know what we see, which is deeply influenced by how we ask the question in the first place. We can predict precious little with the sort of confidence that some would have us believe. Surprises are awaiting us around every corner, and we have so much hard work to do. Progress is needed on so many fronts.

This is why we need safety factors. The safety factor ought to be viewed as a sort of construction sign on the highway of knowledge. It is an invitation to slow down and watch the work needed to keep the road open, safe, and (relatively) free of potholes. We should be more emphatic in educating undergraduate engineers about the stark limits of knowledge so their hubris is levened by an appropriate humility. I first came across the safety factor in learning basic fluid-thermal design for nuclear reactors. Given an analysis of the maximum temperature we would multiply the controlling detail by a factor to make extra-sure that the peak temperature did not cross over the threshold of danger. Another place where safety factors play a key role is the determination of stability and accuracy of numerical calculations. One will compute a time step size that is stable then descrease it by some safety factor to make sure that nonlinearity doesn’t bite you.

Information is the resolution of uncertainty. – Claude Shannon

Most recently there is the case of the estimation of numerical error in calculations. The standard practice is to take the data and produce estimates of the convergence meat-plate-study-1-factor-of-safety-factor-of-safety1rate and converged solution. The distance between your numerical solution and the estimate of the converged solution is the numerical error bar, which is multiplied by a factor that depends upon how hinky the estimates are, the hinkier the estimate, the larger the safety factor. Again we have precise and detailed analysis that is augmented by a factor to account for the slop (well not slop, the nonlinearity). A truism is that if we lived in a linear world, safety factors wouldn’t be necessary. Equally true is that a linear world would be dull as dirt.

That is extremely important. That should be highlighted, underlined and screamed out in an education. We are not masters of the Universe. We are at the mercy of the whims of reality. We only have a tenuous grasp on what reality is. We know far less than we think we do. We have so much to learn. Life is uncertain and fragile. The whole damn thing needs a safety factor. Beyond that, no safety factor will be large enough to save you

 Maturity, one discovers, has everything to do with the acceptance of ‘not knowing.― Mark Z. Danielewski

So, embrace the lack of certainty, and knowledge, it is what makes life worth living. It gives life an edge and excitement. It gives us things to do, knowledge to be discovered, and new truths to revel in. Maybe safety factors are a little bit awesome, and not so pedestrian after all.