In the 1980s software projects were holes down which banks could disappear. The complexity of even small projects took many companies by surprise. Creative teams were ill-disciplined, often delighting in being “artistic”, a euphemism for writing no comments in the code.

A few decades on and the pendulum has swung the other way. Anyone working in the space, nuclear, or aviation industries and living in the straitjacket of “assurance” knows of the frustration of not only taking an age to develop anything but then never being allowed to change it.

How does one strike a balance? The answer may not lie in software at all but in the wider system. After all, the world is full of fallible systems, most walking around on two legs. We get by. Our lives our punctuated with near misses and “you won’t believe what I almost did” moments. The fact is we seem to be able to live with mistakes.

So when Ramsey Faragher, CEO of Focal Point Positioning, writes of trust in systems, he is touching on a much bigger issue than the benefits of deriving position from satellites, Wi-Fi, cellular and inertial technologies. At a deeper level this multi-sensor approach “feels right”. As human animals we seem to be comfortable with the idea of having many senses, even if on occasions we must work out which to believe.

Why the change? One reason is the presence of a new enemy. For decades we have been fighting the “bug”: the mistake made by a careless software engineer one rainy Friday which has survived undetected in the system. We now have other adversaries, most notably malicious people. Spoofing and hacking are facts of life. We now know that residual code “features” are one of many causes of strange system behaviour.

For Focal Point’s wearables and smartphone markets the next few years will see products of amazing quality in terms of reliability, robustness and security. They won’t be fool-proof but they will work brilliantly. This trust derived from multiple sensors is the sort of trust we instinctively understand.

The world of autonomous vehicles will not get such an easy ride. Here assurance still plays its part and standards and attitudes will – quite rightly – not change overnight. Old fashioned assurance wrapped in new-fashioned cyber security will give us more holes down which banks will disappear. If the regulators do their jobs the issue will not be safety but affordability.

Our approach to building safe and secure systems will have to change. The future lies in trust. It’s a much broader issue than safety and security. The pioneering work of companies like Focal Point will find its way in to safety of life systems and with it will come the opportunity to embrace new technologies such as machine learning. It just may take a little time.

Author
Roger McKinlay DipHE MA(Cantab) CEng FIET FRAeS FIoD FRIN
July 2018