Memories of a PDP-8 Wrangler
Our DEC System
I was teaching at a New England prep school when, in the 70's, the math department decided to cut the cable to Dartmouth Kiewit time-sharing and buy our own computer. We contacted a school computing business called Educomp in Hartford, CT and bought from them a PDP-8 to provide BASIC language programming to students on three teletypes, yes teletypes! Two of the TTY's were on the boys' campus in the valley, and one was four miles away and 500 feet higher on the girls' campus.
This original PDP-8 had an RF08 hard drive with 256K 12-bit words (modern equivalent 384 KB), a “high-speed” paper tape reader, and dual DECtape drives. There were 4 sets of magnetic core memory boards, each with 4096 12-bit words.
We had 3 ASR-33 TTY's, each with a very slow paper tape reader/punch. They were noisy! The students only had access to the TTY's. The computer was in the Science building where one TTY was in an outer room with the PDP-8 system visible behind a glass wall.
The whole system cost the school about 10-thousand dollars. I took the processor when it was retired and, for a while, ran it at home just making interesting light patterns. I finally gave it to my nephew.
How It Was Used
The PDP-8 ran the TSS/8 time-sharing operating system which took up about 30% of the hard drive. Users experienced frequent pauses while the system was computing or compiling for another user. Each user got about 4 K of core memory for processing, and some space on the DECtapes to store their programs.
The major use was for students to learn simple BASIC programming, supplementing their math studies. A few enterprising students wrote programs that assigned dining hall tables and student jobs, both of which rotated periodically. The students, and the math, and some science, faculty, were excited about using the computer. The school taught BASIC to the students and then used it mainly for math and science teaching, rather than an end in itself. A math teacher and I ran a summer workshop on BASIC for teachers from other schools that was quite successful.
The PDP-8 system was eventually replaced by little Commodore Pet stand-alones that just did BASIC, and then programming was introduced as a subject. I taught one of the BASIC courses with them.
Most of the time, the system worked pretty well, in between crashes every few days which were usually caused by disk read/write errors that left the computer stuck in an infinite loop. I then had to toggle in the reboot routine and feed the software back onto the hard drive. Occasionally, there was a component failure. The processor used IC's but the disk controller was all discrete components -- transistors, diodes, etc. I narrowed the trouble by swapping out circuit boards. Tedious, but effective. The DECtape drives were pretty reliable, but occasionally, would fail to stop at the beginning or end, and left little bits of tape all over the room from hours of flapping.
There was one really major disaster. The computer was located in the school science building, and the architect, who no longer is used by the school, had the bright idea of putting the building's main power transformer inside the building. All it took was the right lightning strike to burn up 29 components, most of which could be detected by sight and smell.
The other major headache was the teletypes, which were constantly getting out of adjustment, since, unlike newsroom TTY's, they were heavily used in both send and receive modes as well as the ASR-33 being a low cost but less robust model. I spent hours with the maintenance manuals and my feeler gauges and screwdrivers bringing them back on line. In one extreme episode, an angry student decided to bend all the little logic bars because his program wouldn't work. After many days, I did actually get it to go again.
Enter The Business Office
The school business office did all of their student accounts-receivable and their general ledger accounting on a huge mechanical NCR machine that took up half a room. I bought a second PDP-8 system for them with two RK05 cartridge disk drives, with a whopping 2.5 MB capacity each, a video terminal, and a dot-matrix printer with tractor feed. I also hooked up a nifty little rapid entry device for repetitive entries. This looked like a desktop electronic calculator with a small LCD screen. It plugged into the video monitor. I could program it to display just account numbers, followed by money entry, followed by transaction. It made entry much faster than using the terminal directly
Using the Educomp QBOL variant of the business language compiler, COBOL, I wrote programs that did everything the NCR did, but much faster and with better reporting. The programs filled two 3-inch binders -- no Quicken back then. The main frustration in writing the programs was every time I made a change, I had to recompile the program I was working on which could take up to fifteen minutes. Progress was slow. I don't know if this is true, but someone in the programming business told me that I had unknowingly “invented” structured programming based on the modular design I used in the programs. Ah, the fortune I missed, maybe.
One funny thing happened after a few months of successful use. In COBOL, one had to predefine the size of a variable in maximum number of digits stored in that variable. I was told that the trial balance in the general ledger would never be larger than $999,999.99 in any one month, so I gave that variable a size of 8 digits. The first summer, while I was up at our lake home, the book-keeper called me in a panic, explaining that the trial balance came out wrong. Turns out it was over a million dollars. I had her take a pencil and put a “1” in front of the number. She had to do that almost every subsequent month. It would have taken the system offline too long to rebuild the database and recompile all the software with the larger field size, all while the business office was using the system daily.
It took about one school year to write the software around my normal duties as director of audio visual services. The miracle was that when I was all done, except for the trial balance digit, both the student accounts and general ledger ran flawlessly from day one. The business office was ecstatic! I was proud of that, being a rank beginner.
The business office machine ran reliably for several years, printing out bills, checks to vendors and audit-ready reports, and was eventually replaced by an IBM PC.
Sometimes when I get out my iPhone 6S and fire up one of my 26 apps, I ponder how many PDP-8's it would take to do whatever I am doing with it, and how long it would take them to do it. Maybe you, dear reader, can answer that.
A student using the PDP-8 in 1973. The computer is visible over his shoulder.
Part 1 and part 2 of an article about the student PDP-8 system.
The EDUCOMP BASIC language manual
The site managers guide shows the business system was under DEC maintenance from 1982 to 1985. We don't know if that was the entire time the system was in use. We also don't have a record of when the student system was shut down.
If you would like to contribute your story of working with a pdp-8 contact me, David Gesswein email@example.com
Feel free to contact me, David Gesswein firstname.lastname@example.org with any questions, comments on the web site, or if you have related equipment, documentation, software etc. you are willing to part with. I am interested in anything PDP-8 related, computers, peripherals used with them, DEC or third party, or documentation.
PDP-8 Home Page  PDP-8 Site Map  PDP-8 Site Search