One Of The Jobs I Did, That I’m Most Proud Of
In the early 1970s, I developed and attempted to market a simultaneous differential equation solving program called SPEED, which stood for Simple Program for the Efficient Evaluations of Differentials or something like that.
It wasn’t very successful, but two companies asked me to write some for them.
- For Time Sharing, I put the central algorithm into their simultaneous differential equation solver to move the program up to the next level.
- WS Atkins, asked me to install it on their computer.
SPEED had a few advantages over typical simultaneous differential equation solvers of the time like IBM/CSMP.
- IBM/CSMP needed at least an IBM 360/50 computer, which I ran it on at ICI in Welwyn Garden City, where I was a general mathematical dogsbody. But SPEED could run on a dial-up line to a time-shared computer like a PDP-10.
- The time-shared computers as used by Time Sharing and WS Atkins, gave the big advantage, that as the size of the computer increased, the size of the problem, that could be tackled in proportion.
- IBM/CSMP and SPEED both had a simple column-oriented report writer, which unsurprisingly ended up in Artemis, which I wrote a few years later.
- I can’t remember, who at ICI gave me the tip, but I used a sophisticated version of the Runga-Kutta algorithm, that everybody used and some probably still do today. The version, I used was called Runga-Kutta-England, where like me England, was a graduate of Liverpool University.
This summary by Google AI described the algorithm.
The England version of the Runge-Kutta method (developed by R. England) is a highly efficient 4th/5th order embedded numerical integration formula. By reusing intermediate slopes across two methods of differing orders, it provides a highly reliable built-in error estimate for adaptive step-size control.
Does anybody still use Runga-Kutta methods? I suspect not!
A breakthrough of sorts came, when WS Atkins asked me to produce a larger system of the SPEED software, that could handle several tens of thousands of equations.
- Atkin’s client was the Water Resources Board and they were modelling the water distribution system for a large part of the UK.
- Prominent in the project was a Dr. David A, Dimeloe and we became friends and had dinner with our wives a couple of times.
- But I never received a copy of the report, that was written or heard any more since about 1975.
- But judging by the fact, that we seem to have adequate supply of water in the UK and problems seem to be all about sewage, politics, management and finance, I feel that David and his team, must have done a good job.
- I was also never asked to fix any bugs in the software.
I did have some trouble getting the money I was owed from the intermediary in the deal, but I eventually retrieved it through the County Court without a solicitor.
The Anonymous Widower 