Always Have A Pen And....
I read with some interest your Editor's Notebook about the use of the napkin by engineers to solve problems \["Ode To The Paper Napkin—The Untold Story," Aug. 7, p. 54\]. It is true that paper napkins are cheap, ubiquitous, and very useful. I've used them and they sort of work. However, I think that I found a system that works better, at least for me. One of the givens is that I know that I'll need something to write on at some point in the future, for me or for someone else. You know that you're going to need something to write on, so why not plan for it?
I've found nothing beats a 3- by 5-in. unlined index card. It allows fine line drawing (the napkin does not), it isn't hydrophilic (napkins are), and it can be stored and rearranged (try filing a napkin!). And it is cheap. The problem is storage and how you can ensure that you have them with you.
How often do you need to give someone a map that can not fit on the back of a business card? Engineers are always carrying lists. Index cards are perfect. Given a pencil and a couple of index cards, I can work on a To Do list, do a schematic, or list possible approaches to solve a problem anywhere or any time. And I can file and arrange the index cards. As reminders, they are great. And someone else always needs something to write on. Index cards are cheap and I can carry enough of them to get me through a five-day trip.
Compact, light, and always with you, an index card provides a fresh white surface to catch your ideas, carry your work, and store your solutions. All this without batteries. Index cards are flexible, just like napkins.
Use napkins, yes, I guess that I could. But engineers are supposed to plan ahead, aren't we?
Many years ago my employer's CEO, who had been an engineer in a past life, had inspiration hit during lunch. Being a reformed engineer, he whipped out his pen and began sketching on a napkin \["Ode To The Paper Napkin—The Untold Story," Aug. 7, p. 54\]. Being a CEO, the lunch was at a restaurant that was more upscale than those I would normally frequent, and the napkin was linen!
He was outraged when the bill for lunch included a line item for the cost of the napkin. He was insistent that the ink would wash out, but the restaurant was equally insistent that he pay a nominal fee for the napkin. I have long forgotten the nominal charge, and it certainly wouldn't have wreaked havoc with his finances, but the man was furious and created quite a scene. He took the napkin home with him, laundered it, and returned it to the restaurant the next day. He demanded, and got, a refund for the charge!
I, too, have rendered the inspiration of the moment on a napkin, but since that incident I always check the materials of construction first!
I belong to the group that can often go to sleep thinking about a problem and wake up with a solution. I don't keep a notebook by the bed, though. I never wake up in the middle of the night from the shock of inspiration, and the solution is firmly fixed in my mind on awakening.
Energy Strategies Corp.
A Ouija Board Might Work
I have been trying valiantly to explain this "Garbage In Does Not Always Mean Garbage Out" \[July 10, p. 69\] phenomenon to my engineering manager. My point is that centralized scheduling is inherently more fair and more accurate than decentralized scheduling by the virtue of sheer numerical leverage—the inaccuracies tend to cancel each other out.
The trouble is that when centralized scheduling is wrong, the failure is usually spectacular. And, it's usually the result of fundamental changes in practices. How you "divine" these changes of practice is a challenge—I use a Ouija board....
I'm not an engineer, and that's the source of some of the problem, no doubt. I also suspect, as Simon and Garfunkel said, "A man hears what he wants to hear, and disregards the rest." But, I have photocopied your article just in case I hear the creaking sound of a mind opening!
I just read your column "Garbage In Does Not Always Mean Garbage Out" and I have a few comments. First, garbage and uncertainty aren't equivalent. In your examples, any numeric value in the allowable range is uncertain but not garbage. If I submitted an alpha character as an input, it would be garbage and the output would not be useful.
Second, your use of the square root function example and your comment "noise in each component task is never perfectly correlated" reminded me of an article from EDN from March 31, 1983. It proposed treating the time estimate to complete each small task as a mean and standard deviation. When combined, the means add up but the standard deviations combine as the square root of the sum of squares, which results in a smaller number than a traditional addition of all tolerances.
This article assumed that the task times could be represented as normal distributions. This technique also allows you to use the normal distribution tables to determine the probability of completing the project by a specific time. I believe that concept is similar to what you were trying to convey. Thanks for the reminder. Reference "Determine Project Risk Using Statistical Methods" by Rick Palkovic as part of the EDN Design Management series.