My (computer) history 2

1977

Intro

Having being accepted to study Mathematics at St Catharine's College, Cambridge I started an Undergraduate Apprenticeship with British Aircraft Corporation at their Guided Weapons Division in Stevenage.  The first three months was a mandatory workshop training course, mainly learning metalwork, with the traditional apprentices.  It disabused me of any notion that I had any practical abilities and I admired the skills of the younger, traditional apprentices.

After this difficult period I was sent to work in various departments around the site.  The other undergraduate apprentices were due to study Electronics and were easy to find placements.  They didn't really know what to do with a Maths student and it was natural to send me to work in departments with computers.  As a research and development site, there was a significant investment in computers but very little existing expertise in controlling and programming them.

IBM : Fortran : Statistical Analysis

My first placement was with the Precision Products group working alongside experienced traditional mechanical engineers who were computer illiterate.  BAC made extremely accurate gyroscopic compasses for the military.  Repeated readings gave slight variations and the designers hoped to improve accuracy by looking at a statistical analysis of compass readings.

I taught myself the Fortran IV language and wrote programs to analyse readings, using a 80 character coding sheet to write down the program. 

 I then went to a punched card machine where I typed each line of code onto a Hollerith punched card.  Column 1 was blank unless it was a comment.  Columns 2-7 were used for a line number for goto statements. Fortran commands were entered into columns 8-72.  Columns 73-80 were blank.  At the top of each card in each column, the character was printed, and beneath a sequence of holes indicated what the character was in binary. 

 Compass readings were also typed on cards which were appended to the end of the program.

After checking the cards, they were handed to an Operator who used a Remote Job Entry (RJE) terminal to read in the cards and run them in a batch partition on the mainframe.

I believe the computer was an IBM 360 running OS/VS Operating System.  The operator would add a couple of Job control language (JCL) card at the front of my program deck to tell the mainframe which partition to run and which printer should be used for results.

The engineers and myself could look at the printed results.  I also had access to a Calcomp plotter which was attached to the remote mainframe.  This was a very exotic and expensive device which enabled me to show results on a printed graph.

Analog Computing : Missile Guidance

In the 1950s 1960s and early 1970s analog computers were a useful tool.  Signal sources (Input voltages) would pass  through various components to give results in the form of output voltages. They were particularly useful in modelling systems whose behaviour was based on differential equations.

As shown below there were a variety of operations which could be carried out.  BAC used them to test missile guidance systems.  A missile nose cone would be mounted on gimbals in a test rig allowing it to turn and rotate.  A moving Radio Frequency target would be placed at the other end of a long room.   The nose cone then needed to track the target.  Instead of signals send to control rocket motors, the analog computer interpreted them and adjusted nose cone direction appropriately.

Electronic Associates (EAI) were the principal manufacturers of analog computers.  I learnt on one similar to EAI-180 at TNMOC which is shown below. It is programmed by plugging patch cables from inputs, through a sequence of components leading eventually to outputs. Although plugging the cables was easy it was difficult to design/test a working system.  I recall that the production system had a spaghetti-like mass of cables.  Patch boards could be interchanged on the front of the computer to run different "programs".