Chaotic scenes were visible at Heathrow, Gatwick, Manchester and Birmingham airports this week following a glitch in UK air traffic control systems. I mean, I assume the scenes were chaotic and visible. I’ve no idea, because I was on a boat at the time.
The air traffic control outage, the UK’s third in three years, grounded flights for several hours and caused a cascade effect lasting into the next day with some 2,000 flights stuck on the apron.
Authorities came out early to say the collapse was not the result of malicious actors, which is important to note, but what was it then? And what, as they say, ‘lessons will be learned’?
Ordering a rapid report on events, UK transport minister Heidi Alexander flatly stated: “I do not believe this issue was unavoidable”. Alexander will be in possession of information the rest of us will have to wait a while to hear but it’s fair to say she has grounds for saying so. What we, the rest of us, have grounds to say is a broader and, yes, vaguer thing, but no less important: digital fragility is a growing problem.
A column in the Irish Times this week mused about how businesses don’t want to talk to customers anymore, not just because it costs money but because they would, on some level, just like us to pay the bill, thank-you-very-much, and not worry about the service.
Compared to air traffic control going down this is trivial stuff, even if it didn’t feel that way this weekend as I tried to contact my mother’s broadband provider on a phone number that kept asking me why I was calling, an odd question given I was responding to a letter saying they can’t be bothered providing broadband anymore.
The failure mode of digital systems is, basically, collapse. I don’t mean this literally, though it is true of things like digital broadcasting where the bit cliff has replaced graceful signal degradation, so much as I mean systems are designed solely for ordinary situations.
I don’t pretend to know how to make air traffic control more resilient to the stupidity of computers or the carelessness of the people who specify their software, and I accept that with something like massive machines full of people flying through the sky it’s probably better to ground the lot of them than to try and stop them smashing into one another using the kind of associative reasoning I tend to prefer. But there is something worth applying a bit of fuzzy logic to here.
The term ‘fast, cheap and out of control’, made famous by documentary filmmaker Errol Morris, seems to have originated at Nasa’s Jet Propulsion Laboratory (JPL) to describe an engineering philosophy that prioritises rapid prototyping and the cost-effective development of robots over the traditional aerospace approach of slow, methodical planning. Sound familiar?
The limits of digitisation
Metaphorically, it is the story of the Internet and, indeed, personal computing as a whole. The computers we use do not, in any meaningful way, descend from mainframes and minicomputers. Whatever system NATS is using probably does, at least in part, but the machine I am slapping this text into is a jumped-up pocket calculator. In the early days of the public Internet, the 1990s, we all, drunk on the novelty of networks, thought that we could replace faceless institutions with a responsive rhizome. What we ended up with was faceless institutions that act as though they are networks, leaving no central authority, however supercilious and huffy, to interrogate.
What has that got to do with a centralised air traffic control system keeling over, you might ask. In truth, not a great deal on a technical level, but as we fly headlong into more systems thinking, and more networks (metaphorical or otherwise), the fact that computers are only superficially reliable, not to mention dependent on all manner of unglamorous infrastructure, is only going to become more urgent.
Does anyone even ask the question anymore: are there things we shouldn’t even try to digitise? Is the ability to file at speed (which is most of what we actually do with computers, let’s be honest here) causing us to mistake the shape of truth for its content?
As far as reporting goes, the air traffic snafu was, we are told, descended from a technical fault hitting a primary flight processing system, leading to a cascade that resulted in restricted airspace capacity and grounded departures.
It is too much to expect any system, organisation or person to function all the time, and the grounding, while doubtlessly costly and annoying, and likely a disaster for some of those caught in the chaos, was obviously the right response. It would have been more fun for me to respond to the outage with derision and I could have, had I wanted to, bent my argument here to make the derision plausible. It would also have been dishonest. Fast, cheap and out of control, even.
But, more broadly, there is something to say: the digital systems we are lashing ourselves to, from banking to government services, and all of the rest of it, need to be better designed, more flexible so that they can handle so-called ‘edge cases’, and backed up by people, with jobs, who can make decisions and, in extremis, achieve things using another kind of technology: pens and paper.
Whatever turns out to be the proximate cause of the UK’s air traffic hiccup we know human error will be the real issue. Computers don’t think at all. They add up numbers. We can do both.
Ltd
