A dear friend who designs control systems likes to declare, “You have to know the answer before you do the analysis.” He says it with a twinkle in his eye so you know he’s goading you.
But consider his declaration seriously. How does an analyst know he’s (or she’s) right? OK, forget about “right,” how about “good enough?”
The short answer is that he or she may not “know”, but that by developing over time a keen “sense of smell” he or she can “kind of tell”. One way to maintain that sense of smell is to continually make estimates and correlate test data whenever the chance occurs.
In this regard, and with the Holiday Season safely behind us I can pass along an anecdote:
I was managing a mechanical engineering department for a large optics company [the optomechanics key in this missive] and the management club decided to have its Christmas Party on board the Queen Mary, moored in Long Beach. The dinner was preceded by a tour of the ship. My wife and I decided to forgo the tour and arrived just as the diners were being seated.
We joined a small group of program managers at a table near the band-stand. The repartee was reasonably brisk, as it should be among a gathering of alpha-male managers. But after a few minutes the hubbub subsided and I could see that a gentleman across the table who I’ll call Larry had a brochure in his hand. Larry said to me (and I believe I quote him accurately), “Al, you’re so sharp. Tell us [and he gestured to our peers around the table] how many rivets are in the Queen Marry.”
I heard my wife take a deep breath.
Now, Larry and I had some history, of course. Fun stuff like this. And other less fun stuff.
So, I explained to him that I didn’t “know” but would estimate it for him. I estimated the length and breadth of the ship, its number of decks and the sizes of compartments. With that I estimated the length of all the joints needed to be riveted together. Then, estimating the number of rivets per foot on the outside of the ship (which I had observed when day-sailing, with a friend, along her side the previous summer), I declared my estimate: 10,000,000 rivets.
Well, a hush descended on the table that wasn’t lifted until the band started playing. My estimate agreed exactly with the number in the brochure Larry had picked up during the tour. The rest of the table wouldn’t let me buy a drink all evening.
My response to my dear friend is that an engineer who does a lot of analysis needs to stay grounded in the meaning of the numbers by continually making estimates and correlating test data. Yes, even on a pleasant summer afternoon day-sail. I believe he might nod his head and accept that.
Joy to all, including Larry, in the New Year!