25.06.2020

Garbage in, garbage out

Failure of the NHS tracing app is characteristic of the government’s pandemic response, argues Paul Demarty

All told, the UK government’s contact tracing app is one of the more predictable disasters of recent months.

There is, for a start, the fact that it is a large software project with a tight deadline - a situation with an intrinsically high failure rate. As the old cliché goes, ‘On time, on budget, works as expected - pick any two’.

There is also the calibre of the people notionally in charge. (If you want something to maybe show up in September, get Matt Hancock to promise it for mid-May.) As the government lurches from calamity to calamity in its handling of the Covid-19 outbreak, it leaves behind so many broken and spurious promises that it is hard to keep track. Like the doctors in Freud’s dream of Irma’s injection, a parade of ministers is on hand to offer surreal contradictory accounts of how each successive blunder was only understandable and reflected badly on no-one. No minister in my lifetime has ever looked so out of their depth as the feckless Hancock - in his own way as pitiable as a pigeon with a broken foot. It is enough to make you miss Chris Grayling, who was at least a master of failing upwards.

But, above all else, there was the fact that people who knew what they were talking about were warning from the outset that it was never going to work as designed. Not only was the failure predictable, it was - in fact - predicted.

A description of the technical problem may be helpful here. Hancock’s app had two significant components. The first: mobile devices with the app installed and running would use the Bluetooth protocol (a short-range wireless communication method, most commonly used for wireless headphones and the like) to talk to each other. Sufficiently proximate contact would then be recorded by shipping data about both devices to a large central server over the internet.

This is all fine and dandy with phones that use Google’s Android operating system, for the most part. With Apple devices, however, things are different. Apple aggressively defends its phones’ battery power. When the user pockets her phone, most apps she has opened are ‘put to sleep’ - they simply are not permitted to do anything. Critically, you cannot ‘wake’ an app by sending a Bluetooth signal. So the contact is never recorded. The app was trialled on the Isle of Wight, where it was discovered that 75% of Android devices reported successfully, but only 4% of iPhones.

The scandal is not that an approach was tried, and turned out not to work - take it from this programmer: that is just what building software is like. It is that it was known to be impossible from the get-go. The design should have been eliminated before trial. It was no more likely to work than if the engineers had scribbled the code in crayon on a piece of paper and tried to force it into the phone’s charging socket.

So why was public goodwill, time and money squandered on it, to say nothing of the ingenuity and labour of programmers, designers and so on? Perhaps, one day, there will be an interminable public inquiry to get to the bottom of the matter (and we’ll discover that magically it was nobody’s fault). But, in the interim, we may hazard a guess or two. The first - and more creditable - reason is that the only obvious alternative, a contact tracing API (‘application programming interface’ - basically a set of conveniences provided by the developers of platforms - for example, Android and Apple’s iOS - to developers of applications, to make it easy and convenient to use the platform’s capabilities), produced by Apple and Google, is of doubtful use in tracking physical distance at the scale necessary. This is all according to Hancock and Dido Harding, the bureaucrat in charge and former TalkTalk CEO; Harding further promises that her people can compensate for this deficiency, which apparently is Bluetooth-related, and has gone mysteriously unnoticed by the several countries already using the Apple-Google API. So that’s all right then.

The less attractive spin we could put on it is - well - that it is about spin. The UK government’s response to the pandemic as a whole turns out to be an object lesson in the moral and practical absurdity of government by pollster and press-office. The name ‘Cummings’, of course, stands in for all that is nonsensical in this situation: the self-parodically dictatorial ‘advisor’ who invited himself to Sage meetings, and proved unsackable even after his lockdown-busting flight to the north east. Without their philosopher-king, Boris Johnson’s government is merely a ship of fools.

For all the blood-sweat-and-tears Churchillian bluster, the government has been guided throughout by the optics of what is going on. It is more important to be seen to be doing something than to actually do it; hence, for example, the coincidence of catastrophic shortages of personal protective equipment on the front line with airy assurances that huge shipments were about to arrive, or the double- and triple-counting of Covid-19 tests simply in order to meet a meaningless target number that Hancock had essentially pulled out of his arse a few weeks earlier, so as to get out of a different tight spot. It is not Churchill, so much as Brezhnev, that comes to mind.

Prestige-driven

I propose - but cannot prove - that the most harmful incentive in this particular case was the need for a home-grown success that has so long hovered over Johnson and his loyal underlings like a hungry albatross. It seemed the only way to get off the back foot politically. So they promised us a “world-beating” contact-tracing system - leaving aside the implausibility of catching up and overtaking countries like South Korea, which had been doing track-and-trace from the beginnings of the pandemic, building on the experience of Mers and Sars, note that the important thing is that Britain should ‘win’, not that it should actually achieve any real outcome. The human side of this tracing operation has had a dubious career so far (outsourced, inevitably, to one of the major parasitic contractors - in this case Serco - stories continue to leak of its basket-case character). The app, by the same token, needed to be our app; and it is exactly this sort of prestige-driven, inessential requirement that (again, take it from the programmer) leads unfailingly to dreadful, useless software.

We have hammered away at the government on this to emphasis that it is specifically a political, not a bureaucratic, failure. Ultimately we are dealing with the constraints of bourgeois politics - which is to say, the need for rule in the interests of a small minority class to appear acceptable enough to the great majority to avoid rebellion. This is true in ‘democratic’ countries and dictatorships: dictators rapidly fall if their regimes are unbearable for too many, as always happens in the end. The task for the political class (be it a caste of civilian professionals or a layer of military officers) is to present the interests of (a section of) the elite as united with (a large section of) the masses. The problem is that such presentations are always, at a fundamental level, false; and so they tend to move in cycles, as their credibility is exhausted.

In Britain, as in many other countries, we are in the midst of one such transition - from the neoliberal ideology that became hegemonic through its defeats of the labour movement in the 1970s and 1980s to an overtly national-chauvinist Bonapartism. Until March, it was going rather well, from the point of view of the Bonapartists: Johnson cruised to a crushing victory on a Brexit ticket last December, and the stage was set for a brave new world. The transition has been rudely interrupted by the coronavirus pandemic, which demands not the titillating ‘throw away the key’ authoritarianism of Johnson’s Tories, but the inhuman, technocratic authoritarianism which such Tories abhor in the name of ‘British common sense’.

The result is, at least, a short-term setback for Johnson and co, who have been forced into technocratic quarantine measures, but have dragged their feet so consistently that the measures have been ineffective. The result is a worst-of-all-worlds scenario, where enormous economic disruption has been inflicted, but tens of thousands have died unnecessarily anyway due to panic, shortages and incompetence. Needless to say, public trust is at a low ebb, and it is not uninteresting that a second concern, highlighted by critics of the NHS-Harding app, is over privacy, with all data being shipped to some central server. In theory this data is anonymised, but in practice de-anonymisation is commonly easier than people like to think; and, while many might believe that concerns about state snooping are overblown, it is easy to imagine this shower of idiots inadvertently leaking a priceless cache of medical data to persons unknown. (Indeed, Harding herself presided over just such a disaster as CEO of TalkTalk, when up to four million sets of personal and banking details were compromised in a cyber attack.)

The bottom line, as with all the excitement of recent months, is straightforward. Capitalism is terrible at dealing with large-scale public health emergencies: the market must immediately give way to government diktat, government is rendered less effective by the competition between states, and - as we have emphasised here - the legitimacy of government depends on rather intricate prestidigitation, which is easily ruined by unexpected turns of events. Thus the countless avoidable deaths this year, the phantasmatic character of the political response, and the unremitting embarrassment heaped upon poor old Matt Hancock.

paul.demarty@weeklyworker.co.uk