from IBM 305 Programmers

When working for the Army Ordnance Corps in the 50s, I had the opportunity to take an aptitude test for a mysterious thing called a ramac. Fortunately, I passed the test and attended an on-site programming course conducted by IBMer, Bill Brown. This all took place at Benicia Arsenal in Benicia, CA, in 1958. This was the beginning of my wonderful career in data processing.

The system arrived in 1959, and programming our supply application soon began under Bill Brown's leadership. Parallel efforts were underway at other Ordnance Corps arsenals and depots around the country.

We had the pleasant opportunity to rewrite the application a couple of more times before the axe fell on Benicia Arsenal as part of the Kennedy-McNamara base closings. I moved across the river to the Concord Naval Ammunition Depot to program the Univac Solid State 80. From there I went to GSA in SF to work on disk-based IBM 1400 systems, i.e., 1401 and 1410. Anyway, back to the RAMAC.

That was the most exciting time of my career. It was an opportunity to be in on the ground floor of the upcoming computer revolution. The later jobs were good, but my first computer was my best job. Over the years I've found myself thinking how I could solve moderns problem with a 305. Memory now is so abundant and cheap. What I would have given for another few instruction and data tracks!

Bill introduced us to a technique that used the last 20 of the 200 instructions to pull down new instructions to complete a transaction. We would watch the lights flash from 180 to 199 knowing what was happening.

Do you know if any of the RAMACs on display are actually running? If so, would there ever be an opportunity to visit a site and do some programming and board wiring? We used to use scrolls of paper to chart our programs. When working at the console to test and debug, we would unwind those scrolls, which sometimes were very long, indicating that many instruction sets would be invoking instructions 180 to 199.

We had a major upgrade to a dual processor system with dual access arms on the 350s. To prevent simultaneous updates to records we used a feature called dual process interlock and its corollary, dual process interlock suppress.

In one photo on the web, a lady is sitting at the console which supposedly is at Red River Arsenal. I wonder about this as the lady looks so much like a co-worker of mine at Benicia, Marie Wiegand. Of course, it could be that Marie was sent there to help out, but the room looks like what I remember of ours.

Thanks to the start the RAMAC gave me, I ended up moving to the DC area about 50 years ago and retired as head of computer and communications systems at the National Institute of Standards and Technology.

I would appreciate hearing back from you about any 305 systems that may be up and running.

Thanks for your time and attention.

Stephen White - swhite79 @ cox . net

and parts of a second e-mail in response to my comments and questions

I should have mentioned the fun we had developing randomizing formulae. [ Sometimes called hashing functions in modern terminology. Still an interesting challenge ;-) ] Trying to minimize synonyms to keep chaining to a minimum was interesting to say the least. Unfortunately, my best ideas always followed an earlier commitment to a less efficient approach. This was true of my work with both the 350 and the 1405. I got better at it by the time I was working on the 1410 with the 1301 disk drive. You mentioned a pickup truck full of punched cards. That’s what we used to analyze our formulae.

By the way, re 1401 instructions, I recall the fun I had arranging data to use chained instructions. With the right arrangement, all you needed was op codes, which got addresses from registers, or something like that. Many, many years later I was involved in the purchase of a CDC 205 supercomputer, which could operate on data arrays in much the same manner, but several orders of magnitude faster.

My career was fun, but as time went on the politics sometimes overshadowed the enjoyment. As you might imagine, that’s Washington.

I can remember the machine “hanging up” once in a while and the standard procedure was to immediately call the IBM service rep. Coming from an electronics background, I used to feel inclined to pay close attention to what the service guy did in fixing the machine.

One time we got an “Access Interlock” light on the consul. I called the service man and while waiting for him to arrive, I started looking through his service manuals. I was able to guess which bank of tubes was responsible for head access, pull out what I thought was a possible bad tube array and plug in its replacement (which we had on hand for some parts). Long story short, when the service man arrived I had the machine running again. He got upset at that and told me to never do that again.

It was wonderful to bring a friend into the office late at night, turn down the lights and run a program. Pink neon indicator lights would be blinking on the console, the disc access arm was going up and down, moving in and out, the servo relays were clicking away and there was a display light to show off the stack of disks.

Programming was fun too, with both punched cards and a corresponding wiring panel being used for each program.

Nice to discover your web site and to hear there is still interest in such an ancient but sophisticated machine.

Russ Strathdee

email: russ.strathdee@gmail.com
705-722-3966