Question: what is your greates invention?

Guy Rixon answered on 2 Mar 2018:

Bit technical this, please bear with me…

I had to make a control system for an astronomical, digital camera. There were about 20 separate things to do on the cameral to make each picture and the astronomers wanted to make a series of pictures as quickly as possible. Some things could happen at the same time: e.g. the computer could be saving one picture to disc while opening the shutter for the next. Other things had to be done strictly one after the other: e.g. a picture couldn’t be read off the camera until the shutter had closed, and the shutter couldn’t open for the next picture until the previous one had been read off. Telescope time was costing a few thousand pounds an hour back then, so any time saved was valuable.

It’s easy-ish to write a control program that does things correctly and easy-ish to write one that saves time, but actually quite hard to design something that saved time without messing up the sequence and losing some pictures. Even when there’s a design it was quite hard to code it up reliably. Several coder had tried and not got it all right.

I used a Petri net (google it!) to design the program. I didn’t invent that bit, Carl Petri did it long before I was born, but at least I knew to use the notation. My insight was that I could *code* the network simply and easily, in C. I did so and the control program shrank from a few hundred lines to a couple of dozen. It also started working properly and efficiently.

Other may have done this, but I’ve never seen it written up anywhere. I wrote a paper about it and nobody came back saying “this is old, it’s already been done”.

This to me was proper engineering: find a maths technique that works for the problem, design the program with that technique so I knew it was going to work, and then code it in a day rather than a week or a month. Coding doesn’t often work like this.

Sometimes the most pleasing inventions are the small ones that nobody notices…