Rick's b.log - 2022/05/09
It is the 21st of November 2024 You are 18.188.223.120, pleased to meet you!

My first job

I'm only talking about it now because somebody asked me today what my first job was (I said it was so long ago I don't remember, can't have been that interesting) and also because... it's a funny story. If you're a nerd.

I got to sit at a desk in a little office, and basically do sod-all for a wage of around twenty grand a year, in late 1990. That would be about £40K in today's money. As a starting wage.
Like I said, nepotism.

I was overseeing... a group of women. Not secretaries, exactly. There was a really weird setup that the company had two computers, and these women had to transfer the information from one to the other.

Let's start with the first computer. This machine was the "transactor". I don't really know what it did, but I know that it was massive. By massive, think a room about the size of a living/dining room in an average house, lined with big metal cabinets that went from floor to ceiling. It had reel to reel tape units.
My inner geek wet itself in excitement. It's just a shame nobody was allowed to open anything, or I could have squeed at such delights as magnetic core memory and processor boards made of discrete logic and shedloads of transistors. An honest-to-god mainframe of the Big Iron sort.

The newer computer was enclosed in a glass and metal case and mounted on some sort of sprung base in a small refrigerated room. Temperature and humidity were carefully controlled. Three swipe-card locks had to be actived at the same time by different people to go in there.
The machine itself, a minicomputer, was about the size of a household fridge. Vastly more powerful and... vastly less interesting. ☺

Now there was no connection between the two. The equipment was supplied by different vendors and as far as they were concerned, interconnections were simply impossible.

Which led to the women. Their workplace was a room off the main corridor that was partitioned into cubicles. Two smaller rooms were at either end. One end held a big serial printer connected to the mainframe. This periodically spat out information that was, horrifically noisily, bashed into pages of fanfold paper that was twice as wide as a standard page, and had a pattern of alternating lines of plain (white) and green tinted00.

I don't recall what exactly was written, but it was in the form:

1A4E 4E671 503F AB65 349A 902D 9400 981C C461 DEAD C0FF EEEE BABE 091F 7

The women had to enter this gibberish line by line into serial terminals that were connected to the minicomputer. I don't know what the data meant, but I recognised it as being hex.

My job, with my office being the other small room, was supposedly to oversee them to ensure that they were doing everything correctly. But there was little for me to do since the single digit at the end was some sort of check digit. So if stuff didn't match up the line would be rejected and would have to be entered again. This, I would know when one of the troublemakers, who was called Delores or something like that, would shout "you c**t!" and slam her fist on the desk.

Since I was essentially being paid well for doing nothing, some of the women wondered why I was there, until Delores actively agitated to the point where they outright hated me.

On my desk was a tricked out PC. It was an XT, with... was it the 8088 or the 8086? Whichever is the crappy one, it was one of those. An ancient version of DOS, and an ST-506 harddisc that was something like 15 megabytes in capacity. And just to freak out the youngsters reading this, it had a "Turbo" button on the front. Depending on the state of this button, the processor would either run at about 8MHz, or something weird like 4.77MHz. That's megahertz, not gigahertz. And megabytes, not gigabytes. I think the on-board memory might have been 640K (kilobytes).
Inside was a scary maze of ISA cards. The harddisc adaptor, a Hercules video card (that offered a fairly impressive 720×348 resolution, but only in monochrome), and two interface cards that offered a serial port and parallel port each.
Keep in mind the age of this machine. The video card was about a foot long (as long as a school ruler) and had about fifty chips on it. The harddisc card was similar, but if I recall a weird shape, like square with a large piece of a corner clipped off.

On the machine was DOS, Lotus-123, WordStar, and what was probably the original version of TurboPascal.

So I stayed quietly in my office and didn't interact with the bitches women. I would get a stack of completed printouts at the end of the day and pretty much my only job was to count off the first four digits. They ran sequentially, so I checked that each one was crossed out (meaning "done"). The women couldn't lie, they had to sign the paper once it was complete, so if they missed something it would fall on them. So, yeah, I just checked what they had already checked. It was mind-numbingly boring.

...and don't mess with a nerd

The techs would come in at 6pm when the office part closed, and run diagnostics and stuff on the computers until about 9pm. Apparently the mainframe needed nappy changes and regular hugs or it would sulk and not do what it was supposed to do.

I was able to get a look (my eyes only, don't leave the room) at some of the technical documents for the machines. Most of it was not only above my pay grade, but above my understanding (like, uh, a schematic of a logic based maths unit that unfolded to be a piece of paper that would fill a table and was printed so damn small you needed a magnifying glass).
But the part I was interested in was how the machine talked to the printer, in the case of the mainframe, and how the serial terminals worked, in the case of the minicomputer.

Thankfully, the mainframe used something unbelievably slow like 300 baud. The mini was not much better, running at 1200 baud which was apparently designed to allow it to work with a modem.
There was no necessity to enter data in the right order. The minicomputer assembled it, and processed it in one go at the end of the day. So as long as it was all done by the end of the day, everything was fine. It used a very simple terminal that was basically clearing the screen and providing a prompt, followed by a FAULT/GOOD report (literally, that one word). The operator would then press the appropriate capital letter (F or G) to indicate that they had read and understood the message (no hitting Enter to continue) at which point the screen would clear and a new prompt would be shown. Rinse and repeat for seven hours a day. As boring as my job sounded, I didn't envy theirs.

So I tried to get to grips with TurboPascal. It sure wasn't BBC BASIC! But eventually I had something that could receive data in one serial port, and echo it to the other. This was tested by using two spare terminals. Actually, the ones they gave me were VT-something. Decent kit, with pretty screens and text that didn't make your eyes rot. Just something else for the bitches women to hate me for. Oh well, in for a penny, in for a pound.

Without telling management, one of the minicomputer techs created a login on the machine for me to use. The machine would verify the exactitude of what I had entered, and then discard the data.
Another few days of coding, and testing by typing the hex sequences on one terminal and having my PC relay them to the ghost account on the mini as soon as I pressed Enter, everything seemed to be going well.

So after hours, they decided to have a test hooking up the mainframe to the PC instead of me faking it with a terminal. I got some old phone cable and knocked up a quick'n'dirty serial cable to cross the office. I only needed three wires - Tx, Rx, and Ground. Flow control was XON/XOFF, but my PC wasn't a printer, it could keep up.

The test was an absolute adject failure. What was being sent to the mini was complete gibberish. But it was sorted once I asked if there was a way to see what character set the mainframe used. I was aware of this due to using a BBC Micro, RISC OS, and a PC - all of which only had the regular Latin characters being the same. High bit set characters were different. I think on the Beeb they might have just been a copy of the regular set (but Greek or something on the Master range), they were accented characters (mostly) on RISC OS, and under DOS a few accented but mostly box-drawing characters.
The mainframe was something completely gonzo that was not only utterly different, but didn't appear to support lower case letters at all.
Still, I kicked myself for not having thought of this.
The mini used standard ASCII, at least in the low range. I didn't care what the top bit set was, I didn't need it.

The next night, a successful test. But word had leaked to one of the management bods who said that it could not be accepted without a proper transaction log. Well, it was a doddle to spit out each line as it was processed with a timestamp not only to the screen (as I did from the beginning) but also to the parallel port. Bung an old dot matrix onto that, instant transaction log.

Actually, the manager had come in to read us the riot act and threaten to fire everybody, until one of the techs pointed out that they had run through the entire day's transactions and my crappy little PC did something that took eight women an entire day... in about twenty minutes, and most of that was waiting on the mainframe actually sending data.

They brought in somebody from a nearby university to go through the code with me, because none of the techs understood Pascal. He said I didn't really understand it either (which, fair enough, was true) but that I had hacked together something horrible that worked.
Over the next few days, he taught me how to do it in a proper modular fashion, using the correct variable types and "records" (data structures). A lesson that was useful later on when I wanted to teach myself C.

Now a test with a bunch of management people who had showed up. Not only that, but the mini techs had done something to partition off a part of the machine so the data would actually be processed. So it was a live test. Crap, they hadn't mentioned that!

But, our code was good. The entire test ran through without a problem. The techs and various management types looked over the code (I wonder how many understood it) and said that the experiment was a success, and would it be possible to run this system in parallel to compare the results between the daily work of the women, and the PC?
Well, uh, that's what that partitioning stuff was about. It was already being done.

At the end of the week, I handed in my notice. To much rejoicing from Delores who was only too happy to tell me that I'll never do anything useful in life.

I heard, a few months later, that they hired the uni kid as a programmer (and paid him less than they were offering me!) and two weeks after I left, they basically shut down the entire pool of women and just had the computers talking directly to each other. Two of the women were moved to other positions. Delores wasn't one of them. The rest were let go. Bye-bye Delores! <fx: waves>

I was offered a different job, but I declined. By that point I had a job doing "builder's cleans" (that is, cleaning newly constructed houses) which paid a heck of a lot less, but was kind of fun. That saw me through until I got a position being a "Data Processor" for a local company selling property in France. The fancy title was just a fancy title. In plain English, I was the person who typed crap into Foxpro and once in a while prodded the big green button on the photocopier. And made coffee. All the freakin' time. But both of those jobs, and more, are entirely different stories.

The moral of this story? Well, there are two I guess.

1. Be careful getting a job through family contacts unless you are happy to sell out and cash in whilst sitting on your arse doing nothing.
   If you're a mindless numbskull (hint: do you see your job as being incomplete unless it has the word "Manager" in the title?) then it might work out for you. If you have an active imagination and think that sitting doing nothing is awful, no matter what the wage, stay clear.
2. Don't f**k with nerds. Especially don't f**k with nerds that are being paid to do nothing, lest we find creative ways to fill our day, and by creative we mean the sweet sweet feeling of delayed passive-aggressive revenge.

If I'd have stuck it, I'd probably be on a damn good wage by now, and maybe final salary pension. But... jeez... I'm not management material. I have a low tolerance for bullshit, and really don't react well with people who are supposed to "have a clue" but don't. I have to remember not to say anything sarcastic to my current boss's underling when she tells me stuff like I could use Hypofoam as a detartarant (quick explanation: Hypofoam is a Diversey chemical, it's basically industrial strength bleach, which is "mildly" alcaline, like ph11 or something; you use acid to remove tartar; so bleach... would do sod all).

I don't regret walking away. I would rather be a nobody bog-brush wrangler than a somebody who "got there" for reasons other than actual merits and abilities. Especially, given as the two best qualities of my current job are sociable hours, and that I get left the hell alone to do my work by myself. Being a nobody suits my personality.
Sure, fair enough, I don't have loads of cash in the bank and a massive pension pot, but given that I tend to perceive life as a giant cosmic joke, I have at least... had an adventure along the way. Not a nervous breakdown. I'm happy with that.

No regrets.

Your comments:

Rob, 9th May 2022, 22:40

Sounds probably like EBDIC character set, used on early IBM mainframes.   ...and I knew how the story would end as soon as you mentioned you had a PC on your desk...!

David Pilling, 10th May 2022, 02:32

I did something like that - company had stuff on word processors (which were machines not pieces of software), I intercepted the printer output and turned it into data on a PC. I think they decided it was easier to have people retype it.

J.G.Harston, 13th May 2022, 01:58

A friend had a job at a local branch of a government agency that sent thousands of letters out each week. These were printed on Monday, then the staff spent three days putting them into postcode order so they could use the Royal Mail's discount pre-sorted mail service.    My friend noticed you could clone the output, sort it by postcode, then print it. Result: the week's mailing printed out already in postcode order for Royal Mail to collect, in 20 minutes.    Naturally, he got a bollocking from Management, because they were charging HMG for three days' work, and he'd just done it in 1/25th of the time.

J.G.Harston, 13th May 2022, 02:01

I need to get around to typing up the experience of my first post-university job. I think I managed ten weeks. They went bankrupt and were wound up over a decade ago, so it must be safe by now.    Let's just start by saying it was based within the M25, but not in London. ;)