My first task in charge of an officially designated systems team made me responsible for relations between our team of FORTRAN programmers and the company engineers who wanted to use the results of their efforts. This was more of a human relations job to begin with, because relations between the departments couldn't have been worse. I must have achieved something, because the interdepartmental relationship certainly improved beyond recognition, and the company Chief Engineer accompanied me and one of his and one of my staff to witness my presentation of a new project control system I had devised and successfully implemented, to a joint conference of three senior professional bodies, the Operational Research Society (of which I was then a member), the British Computer Society (which invited me to present the paper and which I have since joined) and the Institution of Mechanical Engineers; it also meant a pleasant weekend at the coast in a good hotel at the company's expense!
As this area settled down it gradually became less of a full-time task, and I took on one or two special projects, including the occasional piece of OR work. Probably the most interesting and significant (or so it seemed at the time) special project was to lead a working party with the task of recommending the choice of operating system to use with an impending major upgrade of our IBM 360 mainframe computer. The choice (for a 360/50 with a massive 256k RAM!) was between sticking to what was then called DOS (no relation), or advancing to the new OS, with its multi-tasking environment, spooled printing and other goodies. We chose the latter, despite the need for re-writing all the Job Control Language, program and system testing, etc. this entailed; future developments in the industry certainly proved us right - IBM's mainframe JCL is still basically the same today, plus lots of added goodies.
After that I led a series of projects to develop computerised systems for a whole variety of different departments, (including production control, stores, quality control and finally found myself back on the accounts systems) with varying degrees of success. During the course of this, particularly in work on designing and documenting a production control system for the batch work assembly shop, I became involved in two new and quite different aspects, simply by needing them for the job.
The project proved to be a very big one, so a good system of documentation was needed. Realising this, I arranged to go on a course being run by the National Computing Centre to teach their (then new) documentation system. On my return I adapted this to meet my own insistence on a strictly modular design - if a module would not fit (using standard NCC template symbols for flow charts) on a single sheet of A4 paper it was too big and should be divided into two or more sub-modules to show the detail. This was in 1969, anticipating the better known systems of modular development by quite a bit - the main thing was, it worked, and enabled us to modify the design relatively easily. This led to my having responsibility for training all systems analysts in the documentation system both at Stevenage and Bristol sites of the company
The second thing we developed for this system, again before hearing of anything similar elsewhere, was what has become known as a hierarchical database system to support the systems information needs. Our prime objective, however, had nothing to do with those normally put forward for such systems as DL1, ADABAS, Oracle, etc., which all came along later. Our system was very large, and had to share an IBM 360/50 with a number of other systems. The prime objective, which our design achieved, was faster operation than could have been achieved for the same user system with conventional file handling techniques. We quickly realised the incidental benefits of separating the definitions of data structures from those of the programs using the data, and the two of us mainly involved [where are you Francis Cummins? - if you should read this, please get in touch] went on to develop the specification for a much more advanced approach, which integrated the concept with a form of higher level language.
IBM came out with the earliest version, still on beta testing, of their DL1 database system between our initial specification and programmers starting work on it, but we quickly proved that on our hardware it was impossibly slow and incapable of all the functionality we needed. The company, however, decided it could find the necessary programming resources for no more than the bare outlines of our system, but at least set 4 programmers to work on it for 3 months, at the end of which we had a working very basic database system which was applied in practice to a small number of projects. At least one of these was still in use when I left the company over 20 years later.
Francis Cummins and I seriously considered developing the system privately, but could never find the time to write the necessary assembly language programs nor could we solve the problem of marketing and selling the end product, so after a lot of design effort and numerous discussions we had to abandon the idea. I still have the several volumes of documentation we produced in our own time for what I still believe was technically a winner, and certainly ahead of its time.
Main links within this site:
Jim's Jottings Home Page | An URGENT appeal for help
Running & Jogging for Ordinary People | Jim Fisher Summary Biography | Computers
Genealogy and Family History | Useful General Resources | Gardening with Strange Ideas | Humour?
Investment | Politics and Philosophy | Organisations with a Mission | Science
Wildlife | Map of This Site | Contact me
Top of this page | Jim's Employment Page
This page last updated 5th August 1998