Friday, January 4, 2019

Designing For Failure

     Back in the late Sixties and Early Seventies, when the Space Shuttles were under development, the NASA teams at work on them operated under a peculiar belief. In short, they assumed that testing subassemblies in isolation was unnecessary – that the only level of testing required would be performed upon the fully assembled and integrated Shuttle. This gave rise to a number of calamities, including five major fires. Stubbornness about this approach, which NASA called “Success By Design,” was one of the reasons the Shuttle program took so long to complete.

     There’s information here. How often, really, do human efforts work on the first try? I’ve known a lot of engineers, and I don’t think even one of them would say “Most of them.” (I do know one fellow who claims it’s possible to write software of arbitrary complexity that will work on the first try, but as far as I know he hasn’t yet succeeded at it.)

     Assuming success, even at low levels of component design and fabrication, is a risky proposition. The more components the final product will embed, the riskier it is. A smart engineer tests every element that’s testable, long before he attempts to integrate it with any other element. Sometimes he’ll build special test circuits, good for nothing but testing his subassembly, to assure himself that he’s “in the ballpark” at the very least.

     In other words, a smart, humble engineer assumes not success but failure. Smart, humble people in every walk of life do the same. That’s because they know their limitations.

     Not every walk of life is populated by people who are simultaneously smart enough and humble enough to admit that they have limitations. The one that sticks out most conspicuously is, of course, the world of politics and government.

     Meddling with the lives of others through lawmaking is the most dangerous of all occupations. Very few Acts of any legislature conduce to the ends that were proposed for them a priori. Unexpected costs and unintended consequences abound. Yet you will seldom hear anyone in government, whether elected or appointed, use the Three Little Words that admit to one’s human fallibility: “I was wrong.” You’re more likely to meet a virgin hooker.

     Perhaps we should have expected this all along. The pursuit of political power requires an outsized ego. Such persons are loath to admit to any shortcoming, especially in their intellects or knowledge. They can fake humility when it’s politically expedient, but I wouldn’t bet the mortgage money on ever seeing them practice it.

     Yet a great many Americans still look to the Omnipotent State as the answer to all “problems.” (Here, “problem” should be understood to mean “something that displeases me” to a private citizen, and “something I can base a campaign on” to a politician.) Few are those who understand that what many regard as “problems” are conditions that cannot be “solved” except at an unacceptable cost. Hearken to H. L. Mencken on a subject no politician would say we must tolerate as the least bad of the possible evils, prostitution:

     There is no half-baked ecclesiastic, bawling in his galvanized-iron temple on a suburban lot, who doesn’t know precisely how it ought to be dealt with. There is no fantoddish old suffragette, sworn to get her revenge on man, who hasn’t a sovereign remedy for it. There is not a shyster of a district attorney, ambitious for higher office, who doesn’t offer to dispose of it in a few weeks, given only enough help from the city editors. And yet, by the same token, there is not a man who has honestly studied it and pondered it, bringing sound information to the business, and understanding of its inner difficulties and a clean and analytical mind, who doesn’t believe and hasn’t stated publicly that it is intrinsically and eternally insoluble. For example, Havelock Ellis. His remedy is simply a denial of all remedies. He admits that the disease is bad, but he shows that the medicine is infinitely worse, and so he proposes going back to the plain disease, and advocates bearing it with philosophy, as we bear colds in the head, marriage, the noises of the city, bad cooking and the certainty of death. Man is inherently vile—but he is never so vile as when he is trying to disguise and deny his vileness. No prostitute was ever so costly to a community as a prowling and obscene vice crusader, or as the dubious legislator or prosecuting officer who jumps at such swine pipe.

     If the politician can’t promise to “solve” your “problems” for you, what could he possibly offer you?

     We enter upon 2019 with the federal government divided once more. Many see this as unfortunate, and they could well be correct. But there’s a possibility – granted, not easily measured – that it could be a good one.

     Some things conservatives would dearly love to see are unlikely to get done. A sharp reduction in federal spending is highly unlikely. The dissolution of any Cabinet department, even the utterly useless albatrosses of Energy and Education, simply won’t happen. Few will be the initiatives for which the Trump Administration will secure the cooperation of the Democrat-controlled House of Representatives.

     But many things that would eventuate in disaster are unlikely to get through Congress, as well. A divided Congress could be a barrier to many a big-government boondoggle. It could halt any major expansion of the administrative state, or any major increase in a federal program. The requirement, of course, is that the Republican-controlled Senate must “stand its ground” about such things. That hasn’t always been the case in the past.

     In this connection we must consider the ongoing partial federal-government shutdown as a blessing. No budget? No expansion of the budget, which is a phantasm anyway. No authorization of funds explicitly for the border wall? President Trump has the backing of the electorate, the military, and the Border Patrol, and could find what he needs in pockets of discretionary funding. There are possibilities for improving our foreign policy and our military posture, especially as regards our overseas military commitments, that a divided Congress would find difficult to obstruct.

     Just be braced for failure. For any initiative undertaken over the next two years, be ready with objective evaluation criteria and insist on sunsetting provisions. Remember how seldom any human undertaking succeeds on the first go-round. Let’s pray that President Trump remembers that, too.


sykes.1 said...

I disagree. A divided Congress creates opportunities for compromises. In the past this has amounted to "I get what I want" and "You get what you want". Total spending goes up.

What you are likely to see is either a freeze or outright reduction in defense spending as the House Democrats attempt to transfer defense monies to domestic purpose.

Will Trump trade an aircraft carrier for the wall?

Francis W. Porretto said...

You do understand the meaning of could and might, don't you, Sykes? And what I meant when I said that the GOP must "stand its ground?"

Be braced for failure. I am. But don't be a pessimistic told-you-so type. I get enough of that from my wife.

Manu said...

You can't be a coder and also subscribe to the notion that anything works the first time. When coding, the first attempt is almost guaranteed to fail. If it succeeds *everybody* is usually freaked out.

After you go through a few dozen attempts, you have something kinda-sorta working, but still possessing hidden bugs and problems that probably won't be caught in QA, and will wind up as emergency tickets when the customer actually starts using it.

And, IMHO, political policy is even less likely to work than first-run code.