Google-apper
Hovedmeny

Post a Comment On: Ken Shirriff's blog

"1950's tax preparation: plugboard programming with an IBM 403 Accounting Machine"

17 Comments -

1 – 17 of 17
Blogger bt1138 said...

Another amazing post.

I'm a big fan of your work!

April 15, 2017 at 12:02 PM

Blogger David Cortesi said...

I have a story about the 402. I applied for a job at IBM in San Francisco in the middle 1960s. I had no particular skills for this, other than having spent a couple of boring years at Pac Bell working on switching equipment. Somehow I was routed to the branch office CE manager for an interview, and after chatting for a while, he handed me a 402 counter plate like the one you show above, with the two identical counter wheels and all the latches and pawls and coils. He said, "The one on this side works correctly, the one on the other side has something wrong with it, can you tell me what?"

I had never seen one of these before of course but I sat there across the desk from him and fiddled with them for probably 4-5 minutes and spotted that one small part was missing -- in memory it was a roller, it might have been that cam-follower roller at the 2 o'clock position off the bigger wheel. He says, "that's right, here's the part and a circlip, can you fix it?" and that was easy, so I did.

So I was hired and sent off to Poughkeepsie to learn unit record equipment maintenance.

April 15, 2017 at 3:28 PM

Blogger David Cortesi said...

My wife has an even more relevant yarn from that era. She was an IBM systems engineer supporting unit record customers in the 60s and she got called out one New Year's Eve because the customer had belatedly realized that in the next year the FICA had to be calculated to three decimal places not two. So she says she spent hours deep in the wiring of the payroll board for a 403 trying to figure out how to get one more decimal digit to print. About 3am the janitor comes through emptying wastebaskets and was surprised to find a woman alone in the building messing with the business equipment. But she told him what she was doing and he said, "Oh, well, I guess they had woman riveters in the war, so it's OK" and wandered off.

April 15, 2017 at 3:34 PM

Anonymous Allen Palmer said...

To read this blog is to go back 50+ years to my days as a CE. As I starting reading I thought of the 'insides' of that machine and how as a CE in my dark suite and white shirt & tie we would dive into the internals. The inside could be very dirty and oily ( note the mention of the 'oil pump' in the machine. There was a whole system of small copper tubing that were the oil lines for the moving parts. It took a very heavy duty motor to drive that 2 speed clutch and all the different moving parts. Looking at the machine now makes me wonder 'how did we ever learn to understand and fix the thing'. But learn & fix we did as well as having to understand how to wire and 'read' customer wired boards, You needed to understand the board wiring so you could understand what the customer thought he had wired the machine to do & then if correctly wired why the machine did not perform the job as wired. Young people might look at this type of system as crude but it was not. It took a lot of knowledge, planning and detail to successfully program these machines. It required logical thinking and understand of what the machine was capable of doing. Customers were smart and developed some boards that were so thick with wires that it took a 2x deep cover to fit over the board. I truly believe that it took a higher level of programming thought to wire those machines than it does to program today's computer with it's high level languages and sub-routines. An appreciation of Unit Record unit goes a long way to understand how we got to where we are today.

April 15, 2017 at 4:57 PM

Blogger Unknown said...

Another great article Ken, all I can say is WOW, and congratulations ! Iggy Menendez CHM

April 15, 2017 at 4:59 PM

Anonymous Len said...

Wow, thanks for writing and posting this.

Herbert Grosch talks about plugboards and mechanical calculators a bit in his autobiography "Computer: Bit Slices from a Life":

http://www.columbia.edu/cu/computinghistory/computer.html

April 16, 2017 at 10:06 AM

Blogger Unknown said...

Wow thanks for another great article! Sadly taxes in America have not gotten any easier...

Could you do an article on reverse-engineering the General Instruments AY-3-8500?
It was the core of many pong clone systems in the mid-70s.
(More info at http://www.pong-story.com/gi8500.htm)

Sean Riddle decapped it last February and posted die shots (http://seanriddledecap.blogspot.com/2017_02_01_archive.html)

I'm trying to figure out how it works by myself, but it's not a common circuit like a microprocessor or shift register which is why I'm having trouble with it.

April 17, 2017 at 11:55 AM

Blogger Kaleberg said...

That sure brings back the memories. Granted, by the time I was getting into computing these machines were being phased out. There were still a few of them flopping around in every shop, many still in use. There was a computer language called RPG, report generator, that was nearly inscrutable to any modern programmer, but made perfect sense to anyone used to programming this kind of machine. In some ways, ladder logic is a survival of this type of physically oriented programming.

In "Turing's Cathedral", the author points out that these machines were heavily used in designing the atomic bomb. In fact, the appropriate plug board logic even allowed a form of general programming. A field on each card would indicate an operation between accumulators and counters. This generality, the ability to mix program and data was a guiding force for von Neumann at the Institute for Advanced Studies were he built the first general purpose computer.

P.S. If you ever get around to it, consider the glories of card sorting machines. They were the public image of computing until tape drives came in. They were pretty amazing too. Most of SQL and relational database technology flowed from punch card technology, so in a way punch cards are still with us.

April 18, 2017 at 2:26 PM

Blogger bm said...

David... I joined IBM in 1971 in India. Interestingly during my interview I was also handed 402 counter plate with one of the counters defective; the roller on the cam follower was missing! Appears it was one of the standard tests during that time to test the mechanical dexterity.

March 4, 2018 at 10:50 PM

Blogger Unknown said...

Small correction: the picture directly under the heading "Inside the 403 Accounting Machine" is actually a picture of a 402, because it has alphanumeric typebars on only the left half of the line. The 403 had alpha all the way across.

I learned the basics of 402 programming in High School, as I wanted a chance to work with these dinosaurs. They had a 402, a summary/reproducing punch, a collator,an 082 sorter and an 026 keypunch. Used to run the attendance for all the town schools (about 15 of them) and also to teach with. I worked part time running the attendance.

A couple of stories:
The machines were rented. IBM owned them and charged us depending on how much we used them. We had a 402, because it was cheaper to rent than the faster, full alpha 403. But the guy who ran the unit record room showed me a trick: if you opened the back of the machine, there was a wall of relays. Pull out the right one, and the machine ran twice as fast. IBM lost money when you did this, because you got the speed of a 403, while paying for a 402. They didn't do it too often, but "turbo mode" was nice to have when there was any kind of time crunch.

The next year, the machines were gone and we got an 1130 computer to replace them (along with an 083 sorter and an 029 keypunch. All nice and shiny. The 1132 printer was based on the printing mechanism of a 407 accounting machine. Cheap...and slow.

April 25, 2018 at 7:47 AM

Blogger Unknown said...

Re: 402/403 alphanumerics

I see I have remembered incorrectly (not unusual as I get older, unfortunately) -- both 402 and 403 were only half alpha. The 407 was full alpha.

Carry on.

Face down, 9 edge first, of course.

April 25, 2018 at 8:00 AM

Blogger Unknown said...

I started my career in a school district that was running unit record equipment in 1971. The district operated a 402 accounting machine, 082 sorter and 029 card punch. 3 years later the school district purchased a system 32 and I started my 47 year career as an RPG programmer. System 32, System 34, System 36, AS400 ending this year (2018) on an Iseries.

September 14, 2018 at 2:21 PM

Blogger Kay McClain said...

I'm an 82-year-old retired Systems Analyst. I started my career in the 1955-57 WACS where they taught me how to wire the IBM machines but I never did master the calculator. After the military I ended up Auto-coder programming with one of the first 1401 computers installed on the east coast. That computer only had 4K of memory and programming was no joke in those days. one day we evolved to 8K then 16K ten 32K and WOW look where they are now. It was unbelievable what we could accomplish on just 4K. the old IBM machine plugboards were the forerunners to programming languages and I actually never thought of it that way until I read this article. I see the default e-mail address you use for me is gmail and that is ok. Hope all your readers are having a good day. Best Wishes from Gettysburg, Pa.

May 2, 2019 at 12:59 PM

Anonymous Roy W. Hutchinson said...

I am 92 and started out on IBM with a 45 round hole card are Champlin Refining Company in Enid, OK. I was switched to the IBM "Department" after a girl who had run the equipment during the WAR (WWII). I still play golf and live with my 82 year old beautiful wife in Booker, TX. I still have a copy of one of the 45 column cards along with some 80 column cards. I am a millionaire having inherited the family fortune which my Dad put together when I was a kid growing up in Booker. My Dad was the first elected Mayor of Booker and was the President of the School Board when I graduated from high school at the at of 16. It was some of the men who put the men on the moon who led me to Jesus. And the guy who prayed me into the Kingdom is the only one who worked at Johnson Space Center still works at the Center. I was hugged by Charlie Duke, the 10th man to walk on the moon and was blessed by Pope John in Rome on a trip to Israel. But I wrote a book called God's Trouble Shooter in which I tell of my experiences. I didn't mean to get into all this but Amen!

January 5, 2020 at 10:55 PM

Blogger ananthap said...

To "bm" (ibm in India man).

My company was an IBM customer in Chennai, (then called as MADARS). I attended the last IBM 1401 course conducted by IBM in India.

Where were you a CE?

February 4, 2021 at 4:43 AM

Blogger Bob Flanders said...

Hi ... Thanks for this. In high school (1971) there was a class called "Data Processing", and we had a full set of machines (sorter, interpreter, collator, reproducing punch, and, of course, keypunch machines.) We also had a 1620 where I had an hour a day of time to play (didn't learn much, tho). Our 403 had the letter Q in one column replaced with an + sign so the report cards could be printed. Very fond memories... i have on my keychain the two-holes-next-to-each-other connector used in "programming" these beasts. We would shoot them at each other in home room, and when caught, we had detention. I asked the teacher if I could go play with the 1620, and he agreed. I cam in for "detention" the rest of the year to play with the thing.

May 8, 2022 at 6:28 PM

Blogger wsanders said...

So, each column is represented by one hole, and holes are interconnected by single wires. But each column can have many different “values”. How are the multiple bits per hole turned into a signal that can be carried by one wire? Or are the holes abstract representations of a field of multiple bits?

My dad worked for IBM in the early 60s (alongside Ross Perot!) He used to bring these home for study.

September 8, 2022 at 5:43 PM

You can use some HTML tags, such as <b>, <i>, <a>

Comment moderation has been enabled. All comments must be approved by the blog author.

You will be asked to sign in after submitting your comment.
Please prove you're not a robot