Walt Disney Imagineering 1979
1979 Big Thunder Mountain Railroad Disneyland
So I didn't know it when I interviewed but I was hired to work on the next generation of the Digital Animation Control System (DACS). By the way the name of the software department was DACS, I guess named after the system. DACS was a system that the animators used to create, edit and store the animation data for the Audio-Animatronic figures. Once that was completed the data was transferred to fixed head disks that resided in DACS central at Disneyland or Walt Disney World. This DACS system was a Honeywell 516 minicomputer which filled up most of the computer room at WED.
This second generation system would use the latest Data General minicomputer, the S/250 with much more computing power and slightly smaller in size. I was hired because I had years of Data General hardware and software experience. So I did work on this for a while but got bored because the lead already had the design in his mind and wasn't interested in changing it. To me it was just a polishing of the old methods and that was not interesting to me. To be fair and if he ever reads this, he created a system that the users and my boss liked and this project was consider a success. I did periodically write a few programs for this system as time went on.
But I turned my attention from what was called "Show Control" to "Ride Control". I started tagging along with the two programmers assigned to a new attraction for Disneyland, Big Thunder Mountain Railroad. I learned a lot about how Disney did Ride Control Systems. This is the hardware and software that moves the vehicles around a given attraction efficiently but more importantly safely. Of course one software bug could cause an accident so there was a lot to learn.
The attraction was originally scheduled to open spring of 1979 for the busy summer season but a strike at WED's Mapo division put the production of much of the system behind schedule, and opening was delayed until September 2, 1979.
To date there had never been a hardware failure that caused anything but a downtime on any attraction but it was time to see if Disney could do better. The Electrical Engineer in charge had an idea of a system with redundant computers running the attraction which would constantly compare states of each computer. If they disagreed for more than a fraction of a second, the ride would stop and a message of this disagreement was printed. Thus preventing an out of whack computer causing a safety problem.
The idea was not new to the world but not generally used in most industries, including nuclear power plants. I say this because 2 days after I started was the accident at Three Mile Island (TMI). With what I know now, if the control system for TMI was designed to the Disney ride standards there would not have been an accident. To be clear it is not the redundant computers that would have stopped the accident but the methods and algorithms for monitoring the hardware.
These computers were actually the first generation of Programmable Logic Controllers (PLCs). They were designed to replace relay logic and the target customers were electrical engineers, that sat at a dedicated terminal and pressed keys with electrical relay symbols on them. It was a computer but the language was relay ladder logic. Because of this, the DACS programmers wrote the message generation system in a Data General Nova while the electrical engineer wrote the "software" for the ride itself.
Two of the three previous rollercoasters at this time were in fact controlled by physical relays with the third controlled by a single PLC.
We hired a famous consultant who gave us his thoughts on a design of what was to be called voting. All failures of the system would lead to a particular type of safe shutdown and this voter system would constantly check each computer for their thoughts on these votes. This system was a combination of hardware and software.
Being an Electrical Engineer (and programmer) I volunteered to write this code. So first I wrote a quick assembler/compiler so I could write the code like a programmer and have it translated into the binary code for the PLC. This made writing the code much faster and was able to edit it and recompile it in seconds rather that sitting at that terminal correcting existing ladder logic.
So this was a great beginning, I had only been here a few months and now I had software in a ride.
Back to Disney Years