My blog post, “the Interview” continues to dredge up some great memories, particularly our first day on the job at the test lab at Control Data in Arden Hills. Our crew in Denver had written software to expand the amount of memory accessible to the mainframes to the equivalent of 40mb in today’s terms. Which doubled the amount we could previously use and was a very big deal at the time since the new hardware was already available and our operating system could not yet use it.
Well the software was almost ready to test and my boss approached me, of course at happy hour one Friday night so I was sure to be in a good mood, to travel to Minnesota to run the tests. And she offered me my choice of analysts to make the trip with me on the two week journey. For some reason at the time I thought it sounded like fun so I consented and I told her I wanted my buddy Dan to assist me there. Well Dan had not actually worked on that project, but he was an expert with the operating system and I knew he would be a great asset… So I convinced him, “Come on, it’s only two weeks, we’ll run some tests, go to some Twins games, drink a few beers and we’ll be back home before we know it!”. Reluctantly he agreed and the details for the trip were finalized.
So it was June when we arrived in Arden Hills and checked into the Shorewood for our two week stay. Our first task was of course to go in during the day to check in and get badges and briefings, etc. Then it was off to some dinner, which is when I imagine we discovered the mother lode of free tacos at the hotel bar 🙂 Then it was off to work… Now in those days computers had no permanent memory. Everything that we did was stored on tapes and removable disks, but the removable packs weighed about 20 pounds each. Our first task was to go to the tape library to check out our tapes and disks that we would be using. LOL.. up until now I had forgotten about the trips to the tape library… we would return with probably a dozen tapes worn around our arms like a giant bracelet and a disk pack in each hand. We looked kind of like the Michelin Man carrying all that junk probably a quarter of a mile through the building 🙂
Finally we make it through to the test lab to begin the testing… We step into this giant open room and there are desks strewn everywhere, covered in spare computer parts, wires, terminals, tapes, write rings, manuals and boxes of junk and floor tile pullers. There were cables lying around all over the place, sometimes hooked up on one end to something, sometimes to nothing. And there were also mainframes everywhere, and banks of tape drives and more banks of disk drives the size of small washing machines. And the only instructions we had were that we had mainframe #xx for the night, don’t actually remember the number. And we just stood there looking around in stunned disbelief. To this day I remember my initial highly technical analysis of the situation… As we stood there surveying the mess, I just said “Sh*t!”.
After the initial shock wore off we started poking around and discovered a piece of paper taped to something that sort of looked like a map of the room. We quickly learned that everything had a number and you just had to figure out what numbers went together and you could assemble a functioning computer system. It wasn’t long before we were crawling around on the floor, and under floor with the best of them, hooking and unhooking cables and after a couple hours we had a mainframe, complete with tapes, disk drives and a printer and a Deadstart Panel. Now the Deadstart Panel is an adventure in itself, a panel with a series of up/down thumb switches that are actually the first 12 or 16 instructions that the computer executes, there was no such thing as a BIOS in those days! And it has it’s own map in a manual that you had to look at to set the switches so that the computer can find the channel that the boot disk is on, device numbers and things like that. Kind of feels like you are getting ready to take off in a 747 or something!
Finally we are ready, and we sit down at the mainframe console. The console for those mainframes was the size of a huge old console television and it came with it’s own cabinet on wheels. Right in the middle under the screen was a recessed red button, the deadstart button. It was recessed so you could not accidentally push it and boot the computer. So, Dan was at the console and I was flying co-pilot when the button was pushed. At first, nothing but a “blank tube”, that’s what we called it then when the screen was blank since it was actually a cathode ray tube (CRT) device, and Dan says, “nothing is happening”. I said, “don’t worry, it takes a long time to initialize all this memory”, lol all 40 megabytes of it 🙂 So we sat there for the usual amount of time, and then a bit longer… Still, the “blank tube”… Once again, after sitting there a few more seconds, I offered my highly technical analysis of the situation, “sh*t”. Well there was a way in those days to have the computer barf up it’s memory to the printer, and that’s what we had to do. Hundreds of pages of octal digits, and using the manuals we were somehow able to figure out what had gone wrong.
So as it turned out, we didn’t have this model of mainframe in Denver to test with and the memory addressing was different. A serious oversight 🙁 Our algorithm and methodology were sound, but almost every line of code we had written over the previous year had to be changed to include a variable starting address for the memory. We had counted on it being zero. In that moment, our two week vacation in Minnesota turned into four months of 16 hour nights seven days a week 🙁 By the time we were done, summer had changed to fall and all the leaves in Minnesota were off the trees and raked into piles on the ground. There were many setbacks and a lot more software had to be written, but we finally got the job done and we got to see a lot more Twins games than we had originally planned on, and had acquired a semi interesting story to tell to a future generation of “computer people” 🙂
