Will anyone let me bet my house and pension on this failing?
The UK’s coronavirus contact-tracing app is set to use a different model to the one proposed by Apple and Google, despite concerns raised about privacy and performance.
…
It has opted for a “centralised model” to achieve this – meaning that the matching process, which works out which phones to send alerts to – happens on a computer server.
This contrasts with Apple and Google’s “decentralised” approach – where the matches take place on users’ handsets.
The tech giants believe their effort provides more privacy, as it limits the ability of either the authorities or a hacker to use the computer server logs to track specific individuals and identify their social interactions.
But NHSX believes a centralised system will give it more insight into Covid-19’s spread, and therefore how to evolve the app accordingly.
I’m not really sure what this needs to do beyond what it already does which is tracing and notifying people. Knowing all the cross-matches in a centralised database might tell you number of contacts between people, but so what? How does that make the app better beyond some vague “well, it might”.
The big problem for me is something I try and stress to people which is that if a company like Google, Microsoft, Facebook, Apple or Amazon have already built a service, and you can make it fit what you need, use it. These guys are very good at building huge scale services and keeping them running 24×7. It’ll be better than what you build, more secure than what you can make it and cheaper.
The NHS aren’t. These are people that threw £12bn at a computer system that delivered nothing. These are the people still running on Windows 7 in some places despite it being end-of-life. They’ve had downtime in trusts for days. And they’ve never built systems that handle up to millions of transactions per day, like this will. Are the contractors they use for this going to have the sort of test setups that this requires? They have to design server infrastructure for storage, design and build databases, APIs, as well as the thing that is required with Google/Amazon which the phone app.
I’ll be surprised if this is fully working and useful before this is all over.
Leaving aside cost, privacy and known incompetence, its often forgotten that a major reason companies outsource is to reduce the load on senior management when it comes to day to day operations. Do we really want senior NHS managers being distracted by the responsibility of developing a keep app like this in the middle of a health crisis?
Yes, that’s true. Keep your business focussed on its job. Outsource and just have a couple of people managing the outsource relationship.
NHS and managers are not compatible terms.
Well, first its not a healt crisis, when half the NHS isn’t doing anything, and second, if ti takes their attention away from tryingto run a health service, so much the better.
“I’ll be surprised if this is fully working and useful before this is all over.“
You could have replaced the last five words with “ever”.
well, it’s rather irrelevant if they have a Go Live party the day after we lift all the restrictions.
It’s the usual problem we’ve seen time and time again – and with the PHE refusing to allow private test labs to help. They (NHS management in this case but it’s a wider government problem) can’t allow anything to be outwith their control. Because if it goes wrong they’ll get all the blame without even the usual illusory control. Part of Chris Dillow’s managerialism rant (but that is wider again.)
You might be right, although this is still outsourced to a company to build. My other observation is that people specify every little thing they want, rather than accepting the often trivial limitations of what’s off-the-shelf. Not precisely what we want, so we’ll build it.
The private sector generally compromises more on these things and understands that tapping into an existing software or process means you don’t get snags and it’s far cheaper.
It’s already all over, all we have now is the tail end of it. Deaths peaked on 8th April, so infections peaked some 10-20 days before that. With a bit of luck, and an absence of cheating, the decline in deaths will go exponential soon. Then we twiddle our thumbs for a few weeks more, then we waste yet more time getting back to normal and then we realise we have been frightened of our own shadow. Then we turn on the government.
Deaths will decline because attrition in the group where mortality takes place means the virus is running out of victims, and the remainder are harder to find. It’s maths, not lockdown. The whole population was never at risk of mortality, only about 1%.
Your former Register colleague, Andrew Orlowski, had a piece in the Telegraph on Saturday covering this general topic. He doesn’t give the NHS app much chance, either.
Covid tracking. When we have the answer, what question will it answer?
It is a fast spreading virus, which will in time infect more than 50% of the population., possibly pass 70% and give herd immunity.
It has already caused disease and/or killed those most susceptible.