Networking the computers turned out to be both easier and more difficult than I had imagined. Windows for Workgroups 3.11 made it easier than I had perhaps thought, and cheaper too. Mid 90s networking hardware made it ridiculously error prone, difficult, and liable to fail at the worst moments.
The word “Network” typically conjurs up an image of many computers connected together. The company I worked for owned one set of Microsoft Windows floppy disks.
Floppy disks (for those who entered the work arena after the late 1990s) were the main method of software delivery before CDs became cheap in the late 90s and early 2000s. They are interesting from a historical perspective because they were categorically not floppy, and held an enormous 1500Kb of data - about a 400th the amount of a CDROM. Nerds reading this will be holding their hand up, shouting “what about 5.25C/3 floppies?” (which really were floppy), and bearded Unix and CPM admins will be shouting “I remember 8” floppiesother options.
First option was Microsoft Disk Operating System - known the world over as MS-DOS. DOS applications were stable, ran like lightning, but didn’t play well with others - you could only run one thing at a time. There was a competing, compatible system called DRDOS, which ran faster, and was more efficient, but they eventually got sat on by the giant from Redmond (who we must remember at this point was only a playground bully - not yet the all conquering marketing machine we now know).
The second mainstream option was Microsoft Windows 3.1. , installed on countless home computers the world over. Windows provided a graphical interface to your computer and ran on top of MS-DOS. It allowed the magical ability to run more than one application at once within “Windows”. It had a slightly bigger brother called “Windows for Workgroups 3.11”- which in the bizarre naming scheme attributed to Windows, and continued to this day, may as well have been called “Windows 3.5” - because it was.
Workgroups made sharing files and printers much easier than it would otherwise have been.
Another option that also existed was IBM OS/2. Once upon a time IBM and Microsoft worked together on a new 32 bit operating system that was to be “the future”. It was a brave, and bold move. eventually killed in the same way the early Mac had been; Microsoft fell out with IBM - but not before secretly building their own version of OS/2 - Windows. The crying shame was that OS/2 was years ahead of Windows in just about all respects.
Striking parallels with the great video format wars of the 1980s, Microsoft spent their money on marketing rather than making Windows better, so therefore won. The better product lost - to the detriment of us all. Granted, OS/2 was massively more expensive, and none of your old software could be guaranteed to work, but it promised true 32 bit computing. In laymans terms this is similar to the supercar debate. Most computers are capable of far more than they ever achieve - similar to using a Ferrari to collect the shopping. OS/2 provided a conduit through which the hardware could be set free.
Meanwhile, we all used Microsoft Windows like the fools that we were.
Behind the entire Windows gloss and sheen, we still had the same computers we had been using since the late 1980s. The architecture was still being dictated by IBM, and perhaps most importantly, USB had not been invented.
We take “plug and play” for granted now. In 1995, it was a pipe dream. Connecting a computer to a network involved taking the system unit to pieces, inserting new circuit boards, and physically configuring wiring and settings on the boards for the computer to once again work correctly (which they frequently didn’t). Any hardware you added within the beige box would argue with anything it possible could - forcing you to remove the lid and tinker some more.
The operating system didn’t “see things” either. You not only had to install drivers to educate Windows how to talk to a device, you had to tell it where the device was in it’s imaginary world - as configured by the wires and jumpers you just monkeyed with. These mysterious settings were called Interrupt Requests, and Addresses, and they were known to drive sane system administrators up the wall.
Networking computers was another layer of luck, complexity and buggeration.
In order for two computers to talk to each other, they both needed “network cards”. The standard back in the day - and still today - was Ethernet. Granted, it wasn’t as fast as now, but it was essentially the same technology that was invented off the back of the Arpanet many moons previously, which eventually became what we now know as “The Internet”.
The world owes an awful lot to two gentlemen - Vint Cerf, and Bob Caan. I’ll let you look them up.
One of the more colourful stories from the advent of ethernet comes from the lab where the first implementation of was developed. It was late because the genius engineer working on the circuitboard got bored with what he was doing, and thought it might be a good idea to start testing the sound reflectivity of the ceiling tiles in his office. This story may or may not be an urban myth, but it wouldn’t surprise me at all.
On a wicked parallel (and rat hole), our new office starter, Halley, eventually lost her job because she started filing work she found boring underneath her desk. The papers were discovered during the christmas break, and I found myself wondering why she was in tears after I wished her a happy new year… turns out she was emptying her desk.
Putting a network card in an early 1990s PC was fun. You not only had to buy a card that would fit the internal connectors (there were several different shapes and sizes); you also had to pray that the constructor of the machine had left you enough IRQ or address space to configure the damn thing. If all went well, the machine would start up, the driver software would load, and the light on the back of the box would start blinking - saying a repeated “hello?” to the unknown wider world. If you were unlucky, you would hear several loud beep noises. Computer language for “you’re shagged”.
Ensuring a computer still had a sensible quantity of useable memory left after starting was akin to the power-up procedure agonised over by Ken Mattingley aboard the ground simulator of the Apollo 13 capsule. Memory management was a black art I became particularly adept at - and quite possibly the most boring subject imaginable to impress people with at parties - not that you ever got invited to any parties of course…
With two computers sat on my desk - one at either end - and all the requisite 10BaseT connectors in place, I started crossing fingers, touching wood, and praying to various deities.
It worked!
If you had everything just so - which it appeared I had done by blind luck - the computers could see each other. The next few minutes were filled with air punching, and demonstrations of copying files back and forth for anybody who would take an interest.
“Why would we want to do that?”
Sian didn’t look convinced. She got the forced demo while collecting mugs to make a round of coffees.
“It would mean Sam could open the letters Halley writes and approve them.”
“But I can print it out and give it to him”
“This is quicker though - look - bang - it’s there”
“I don’t know - by the time I’ve figured out how to do all that, I could have printed it and walked down here.”
I knew I was on a hiding to nothing.
Email and shared calendars had not become the selling points of networking yet. In 1995 very few people had internet connections at work or home - let alone email accounts. Of course I had one, but I was the only person in the real world that I knew with one.
Early steps on bulletin boards - the predecessor of social networks - had opened my eyes to the concept of “real” and “internet” friends. I counted lots of people all over the world as “internet” friends, and yet knew deep down that I would never meet them face to face.
Having figured out the basics, I started looking in an office supply catalogue for ethernet cable. We were going to need kilometres of the stuff. Looking around the office, I had no clue where we would hide it either. In reality, it would never be concealed or hidden - doomed to lay underneath desks for the duration off the company’s existence.
Given that any broken point on a BNC network cable causes the entire network to fail, you can imagine the hilarity that ensued when people stood chairs or tables on it.
“Jon! Damon has crashed again…”
“No he hasn’t - it’s working fine on my computer here.”
“Well mine isn’t. Come on - pull your finger out - we’re losing money here!”
“Okay”
I then set about following the cable up the office - more often than not finding that George or Darren had moved their desk or filing cabinet and broken the cable. They never learned.
In these days of enterprise grade open source database servers such as MySQL, it’s easy to forget where we came from - a world inhabited by the rumoured powerhouse that was Foxpro (which nobody could afford), and Microsoft Access - which Microsoft had built out of spite to destroy the fortunes of Ashton Tate when they wouldn’t sell them Foxpro.
Of course, Microsoft later did buy FoxPro - at a vastly reduced price - and eventually threw it away.
Probably not a good idea to say no to Microsoft.
Running across our Heath Robinson network, Microsoft Access worked (unbelievably), and let many people look at the same data at the same time as each other. Unfortunately any sign of difficulty along the inter-computer telegraph wires caused it to go through an internal mailaise similar to a brain tumor.
Sure, the tumor could be removed, but it meant taking the database away to database hospital (my computer) for brain surgery - meaning nobody could use it until it was all better.
In the end we narrowed down the causal factors for data corruption to snails farts, how the weather felt, and which direction Worzel Gummidge had fallen in the closing credits of his show last weekend.