Adventures in Fresno

After working on a few projects, I was assigned to write an error-recovery routine for one of the first automated teller machines, by Diebold, which had been custom-built by them for Wells Fargo Bank.

The first ATM was actually invented in 1967, but as of 1980 I'd never heard of them. I'd also never heard of Wells Fargo Bank, though of course I knew about Wells Fargo stagecoaches, as I grew up in that decade where every TV show was a Western. But Wells Fargo Bank, in 1980, really only had branches in California, a place I'd never been.

After spending some months working on the Tandem software that would ensure no computer outage could cause money to appear or disappear from anyone's account, I did get to visit California for the first time, a town called Fresno that was so small, their airport couldn't accommodate jets. So I flew from Omaha to Los Angeles by jet, and from there to Fresno on a small passenger jet.

My seat was next to that occupied by a kid whose mother and sister sat behind us. But the mother was sound asleep. We hadn't even taken off yet, and she was dead to the world, leaving the little girl next to her to kick the back of my seat.

Worse, the little boy needed help with his seatbelt. After I fastened it around his waist, he said, "I want it on tighter."

"Really? It seems snug enough, but okay." I gave the end of his seatbelt an extra tug.

"I want it on tighter!" the kid insisted, urgency creeping into his voice.

Shaking my head, I shrugged and gave the belt a really good tug.

Now almost crying, the kid said, "I want it on tighter!"

"It won't go any tighter without cutting you in half," I explained. Then the light dawned. "Oh! You mean un-tighter!" And I loosened his belt, and that was the end of it.

Our installation of the new software did not go smoothly, and we had to stay another week. But we had the weekend off, and I took advantage of that to fly to San Francisco to visit my co-op friends Dick and Rick and camp with them in Yosemite. But then I had to fly the little plane back to Fresno, this time from San Francisco. Nearly everyone was on board but we had to wait for two passengers. I saw them as they emerged from a van in front of the aircraft door steps. It was a morbidly obese woman and her equally obese daughter. The plane groaned when the woman planted her foot onto the first step. When her daughter joined her, the plane dipped toward that side.

And their seats were on the same side of the plane.

After we took off, I was conscious of the pilot's having to trim the wings or whatever it is they do to try to keep level when the plane isn't evenly loaded. Which it would have been if they'd simply put the woman and her daughter on opposite sides of the aisle. But apparently no one thought of that.

Luckily, it was a fairly short flight, just a little over an hour. But as the plane nosed down to lose altitude, the obese woman started wailing at the top of her lungs, "Oh, holy Jesus, we're all gonna die!" She kept screaming these, and rocked back and forth in her seat, while her daughter rocked with her, trying to calm her down, saying, "Now, Mama!" Their rocking from side to side literally made the plane, which was trying to land, rock from side to side. And I thought, She might be right, we could die!

The landing wasn't the smoothest, as we hit one wheel before the other and bounced a little bit before coming to a stop. There was a breathless silence, and then the mother spoke up. "Well, I hope I didn't upset anybody!"

This trip was also the first time (of two) in my life that I've gotten drunk. (And yes, I was 30 years old at this point.)

When I was in high school, there was plenty of peer pressure to smoke cigarettes and drink alcohol. But peer pressure never worked on me. The more people tried to get me to do stuff, the more resistant I got. And so I never smoked, and I never had more than one cocktail.

But on our installation in Fresno, hanging out at the pool after hours with my coworkers Mike and Ike (their real names, I swear!), one afternoon Ike expressed a desire to "get drunk". He was normally a very sober, if sociable, fellow. And for some reason, I decided that this would be the day I'd find out what all the shooting was about, and get drunk with him.

It seemed like a safe experiment. We were in the hotel; we wouldn't be driving anywhere; and there was no family present to worry or judge. So I went halves with Ike on a fifth of bourbon.

We had a couple of drinks, and Ike decided he'd had enough and went back to his room. But I didn't feel drunk yet. So I had a couple more.

Suddenly, I got very sleepy but also the room seemed to be spinning around me. This continued even after I laid in my bed. I had heard the term "bedspins" but now I knew that that meant.

I didn't fight it, but it was most unpleasant. The only previous time I'd experienced anything like it was when I was 9 or 10 and my sisters and I spent the entire day seesawing. When I went to bed, it seemed like the room was moving up and down.

There were two reasons I chose bourbon. The first was that I had had a sip or two of bourbon previously, and it didn't taste horrible. The second was Ike's advice: "Get drunk on bourbon and you'll pop out of bed in the morning.

When my alarm went off at 6 AM, I popped out of bed, all right. I popped right into the bathroom where I proceeded to puke all my insides out.

When Ike and Mike showed up to get me for work, I was still sick but managed to accompany them back to work. No breakfast, but I was able to eat lunch. What I still haven't figured out, is how anyone gets drunk a second time.

(Okay, I got drunk a second time. Not intentionally, but because I hadn't learned how much I could drink without getting sloshed. But I definitely can't understand anyone intentionally getting drunk a third time!)

But speaking of having lunch with Ike and Mike, it was also on this trip that I ate (and fell in love with) fried shrimp for the first time.

My sisters had eaten shrimp when we were kids, but I was stubborn (even though I had to admit it smelled great—but then so does coffee and oatmeal, and I didn't like those, either) and had never actually tasted it.

But Ike and Mike chose a Long John Silver's drive-thru for lunch; and when both of them ordered fried shrimp I impulsively made it three.

And—oh my gods!—it was the best thing I'd ever tasted! I couldn't get enough of it, and ate shrimp for lunch and dinner the rest of the duration.

That was all the fun stuff I managed to work into the installation trip. But it wasn't all fun.

Our manager, Earl, was not computer savvy. (This was not unusual, by the way. I've had a number of managers who personified the Peter Principal in that he was, I assume, a perfectly normal guy who had risen to the level of his incompetence.) Having been promoted to a position for which he was ill-suited, he would never be demoted, but would never be further promoted, either. So there he was, in middle management, trying to manage a computer project he could never understand.

What he did do well was placate and coddle the customer, by telling Wells Fargo what they wanted to hear: The project was going well, it would be done on time, we would only need one week to install and test it, and they'd be good to go.

Except, none of that was true. Diebold was still updating the software on their new model ATMs, so code that worked this week might well break next week. My portion of the project, besides writing the basic error-trapping code for the ATMs themselves, was a module called Store-And-Forward. I was able to modify a store-and-forward from a previous project; so I spent most of my time helping Ike and Mike with their modules.

Okay, so let me briefly explain how the system worked. See, before there were ATMs, people conducted their banking business by physically going to the building housing their bank, where people called tellers (the non-automatic kind) would accept deposits and issue withdrawals, putting it all in a copious paper trail that would then be typed into the mainframe computer by people called "keypunch operators". But the transaction wasn't actually applied to the account until some time around midnight, when a "batch operation" would be started for the day's transactions. It wasn't unusual for the mainframe itself or the batch processing program to crash, so a backup would be applied, the cause of the problem found and fixed, and the batch restarted.

Obviously this wouldn't work for ATMs. What if the mainframe crashed just as someone made a deposit? The machine would swallow the customer's money, which the customer wouldn't stand for. Or, what if the mainframe crashed just as someone made a withdrawal? Or if the ATM crashed trying to dispense the cash? If the customer got cash without it being noted, that was something the bank wouldn't stand for.

Luckily, Tandem computers came to the rescue. They had a dual-CPU architecture in which each CPU ran the same program with the same data, so if either CPU failed, the other would continue to run as if nothing bad had happened. Obviously, we had to write hooks in our programs to keep the two copies in sync; and we had to include lots of error checking at every step of the way.

So, as you can see from the above graphic, the Tandem stood between the ATM and the bank's mainframe. When the customer used the ATM, the transaction went to the Tandem (and my Store-And-Forward module) which, if the mainframe was "up", would pass on the transaction instantly. But if the mainframe was down (or unavailable for any reason), the Tandem would simply store the transaction until the mainframe was back up, then forward the transaction to the mainframe.

The Tandem also maintained a copy of the bank's customer account database. Not the whole thing, but enough so the ATM could report on balances and allow withdrawals even if the mainframe wasn't running.

So our programs had to work, reliably, every time. Yet TAL (the Tandem Application Language) is a very low-level language; and the TAL compiler itself wasn't without flaws. For example, when Ike had a problem with one of his modules, and neither of us could find the issue with normal debugging techniques, I flipped through the program listing and stopped on a particular page, for no particular reason. There was a variable definition and assignment, and, while it looked legal and produced no errors, nevertheless I just didn't like the way it looked. Since there were other ways to declare and initialize a variable, I just changed the way the code was written, not the underlying logic. Nevertheless, when we recompiled the code and tested it, the original problem did not recur.

Ike and Mike both looked at me kind of funny after that. But they always came to me first when a bug needed extermination. (Ironic, since I'd been a pest control person almost a decade before.)

As the installation date approached, it was obvious to the three of us that we would not be ready in time. So, we told Earl. Earl did not tell the customer, or his boss, this important fact. And so we flew to Fresno for our "one week install" that took two weeks, and then a third! The customer was not happy, and complained, not to Earl, but to Jim Cody, the boss. But Jim didn't ask us what had happened; he asked Earl. And Earl, as they say, threw us under the bus. He claimed we'd never reported to him anything but that the project was going swimmingly. So, obviously, we were all incompetent, and that's why the project almost failed.

But Jim asked Ike, Mike, and myself to write up a report explaining what the problems had been, and how we could avoid them in the future. And I wrote one, and I did not hold back.

Sadly, I no longer have a copy of my essay. I do remember I titled it, "How I Spent My Summer Vacation". I pointed out the problems that had arisen, which wouldn't have been problems if we'd had adequate time to iron them out. I described, with dates, the problems I and the others had reported to Earl, and his non-response other than to assure the customer we'd be done on schedule. I recall using the phrase "sycophantic sophistry" to describe his dishonesty. In conclusion, where I was supposed to recommend a solution for avoiding such a situation in the future, I was blunt: "Fire Earl. We shouldn't delay his future success at the window of a fast-food drive-thru."

Then I was invited into the office of Norm, the vice president of the company. He began with, "We've decided not to fire you."

"Fire me?" I laughed. "If you think telling you the truth is grounds for dismissal, never mind; I quit."

"Not so fast," Norm countered. "I know perfectly well that your description of the events is accurate. But we can't fire Earl for reasons that are neither here nor there, and he demands you be fired. But you really understand this stuff, and you are a clear communicator—if maybe a little too clear. So we've decided to put you in charge of training classes that we give to the programmers of our customers so they'll understand how to use our products. There won't be a raise, it's a lateral transfer, but your hours will be a simple 9-to-5."

I blinked, shrugged, and accepted.

And that's how I began my career as computer programming instructor.