When Not to Automate

by

Old Tom

Otscripts

You have a computer, or you wouldn't be here. Use to your benefit! If you don't already have the right tool for the job, download it or order it online. No matter what you do online, there's a tool available to do it for you.

Given that much of my web income comes from script sales, I obviously have a vested interest in suggesting automation. Even so... Don't! There's a lot more money to be had in doing it by hand.

I noticed something at the convenience store last night. Sandy made a purchase of $30.73. I forget the precise amount, but it was between $30 and $31. She handed the cashier a hundred-dollar bill and a one-dollar bill to pay for the purchase. He had absolutely no idea of how to handle it. He finally entered $1100.00 into the electronic cash register. At least he recognized that $970.27 was not the right change... He didn't even know that Sandy is the store manager, but no doubt that will become clear soon enough.

I see this sort of thing quite often, that people can't make change. For a purchase of, say, $11.23, I'll hand over a twenty, which is fine. But when I also lay down a dollar and a quarter, I have hopelessly confused the issue. Often as not the cashier will ring in the $20, ignoring the $1.25, and carefully follow the cash register's instructions for giving me my change. As they're doing so, they'll inform the manager that they're running low on one-dollar bills. (I trust you see the connection.)

Time and time again, I'll see high-school-age people - honor students - incapable of making change by hand. Our particular school district, obviously, teaches the kids how to do arithmetic *on a calculator*. It seems to me that a necessary skill is lacking here... but this essay is *not* a rant against our local school curriculum.

Let me give another example.

Years ago, I taught operating system internals to our field maintenance people. Much of the work involved converting numbers between octal and decimal, but also adding, multiplying, etc., in octal. We all had calculators which will do this perfectly; and, the computer we were studying had various tools to do the interpretation for us.

However, I absolutely required that my students *not* use their calculators in class. I showed all calculations on the board, and I was clearly *doing* the calculation by hand, on the spot.

With continuous and privileged access to a ten million dollar supercomputer designed for number crunching - in octal - we sat there with pencil and paper. Time after time we took half an hour to do what should have taken less than a moment. Every class, someone asked the obvious question: Why? Why slow down the problem analysis? Why make things slower, when the basic tools are at hand?

Now we come to the point of this essay.

Remember Murphy's Law: "If something can go wrong, it will." (Look up the origin of this Law sometime; it's an interesting story.) In brief, the designers assumed that if something *could* go wrong, that it eventually *would* go wrong. The result was an outstanding safety record. People survived.

Meanwhile, we're standing in front of this class of perfectly competent people deprived of their calculators.

I asked them if they could picture getting called in to work at 3am because the computer was down. These were experienced field maintenance personnel; they could picture the situation all too well. I asked if they could picture still being a bit sleepy as they headed out the door, and perhaps forgetting that they'd brought the calculator home to add up the checkbook. Yes, they agreed, that *could* happen.

Could they picture arriving at the computer room, to discover customer personnel already there, wanting to know why their precious multimillion dollar computer was down, and when it would be back up. Could they picture the arrival of the Director of Operations, the Vice President, and soon the Executive Vice President. Yes; several had had similar experiences.

I asked them to then picture explaining to the Executive Vice President that the computer was going to sit there for another hour before *anything* happened. And why was this? Because you left your calculator at home, and were completely incapable of doing your job without it. Or, better yet, it would *stay* down until 8am when the drugstore opened and you could buy new batteries.

With that picture firmly in mind, every single student chose to learn to do it by hand. And, having done it by hand, I'm quite certain each one was *very* sure they always had that calculator handy! Obviously every person had agreed that learning to do it by hand was better than having to call for help because they didn't know their stuff.

The tools are great, when everything's working. You can get a whole lot accomplished, perhaps without even lifting a finger. Crontab, autosubmit, let the good times roll. If something tells you it's broken, no problem; get help and get it fixed.

But... but... what if you don't have a clue? What if something is broken, and you don't know it? Remember that the whole point to automating is that you can set it up, and forget about it. But... what if you're clueless?

That's why it's critical to *your* business to do it by hand first. It's *your* business, and it's critical that you have a clue. If something can go wrong, it *will*. Plan on it.

If you choose to become expert by doing it by hand first, you'll instinctively know what to do when things go wrong. You'll *know*. Furthermore, if something *is* going wrong, there's a much better chance of your noticing - of your recognizing that something isn't right. Finally, you'll know what to automate, and when. You'll know how to make the best use of the more powerful tools, because you *understand*.

You understand, because you did the whole thing by hand first.