Home › Forums › Chat Forum › The contact-tracing app, accuracy?
- This topic has 518 replies, 138 voices, and was last updated 4 years ago by Greybeard.
-
The contact-tracing app, accuracy?
-
JamzeFull Member
So Hancock has jus said the failure of the app is not their fault but Apple’s?!? You mean the issue that everyone flagged up from the beginning?
Unbelievable the brass neck he has.
Not quite.
As we all knew, our own app won’t work well on Apple devices.
What they’ve discovered in testing is the Google/Apple version of the app (the one everyone says we should be using) doesn’t work well enough either. Our app is actually better at the contact detection bit as it calculates distance better.
So we’re now working with Apple/Google on improving their solution – although Apple seems unaware 🙂
DrJFull MemberWhat they’ve discovered in testing is the Google/Apple version of the app (the one everyone says we should be using) doesn’t work well enough either. Our app is actually better at the contact detection bit as it calculates distance better.
Worth noting that Apple and Google don’t actually make an app.
JamzeFull MemberHow much truth is there in the rumour that the developers were pals of Cummings?
Or was it just the horrible **** at Serco?
I think this was debunked. Pivotal are the actual developers of the app (acquired by VMware last year). I assume Serco provide another element of the service.
https://fullfact.org/online/not-dominic-cummings-sister/
https://www.vmware.com/radius/pat-gelsinger-contact-tracing-apps/Worth noting that Apple and Google don’t actually make an app.
Agreed, it’s an API, our app sits on top. The Apple documentation does mention a bare-bones standalone app will be released in coming months, not sure how that would integrate with the health service processes.
CountZeroFull MemberWhat they’ve discovered in testing is the Google/Apple version of the app (the one everyone says we should be using) doesn’t work well enough either. Our app is actually better at the contact detection bit as it calculates distance better.
So we’re now working with Apple/Google on improving their solution – although Apple seems unaware
On This Morning right now, Apple have had no contact or discussion at all about this.
Hancock is openly blaming Apple, completely ignoring the fact that this app is designed to be platform-agnostic, able to function across a whole range of makes and models on iOS and Android! ****wit! 🤬
Yet another example of why our government should never be allowed to develop large IT systems, because they’re completely ****ing incompetent!
Just like that cluster**** NHS system that the Labour government under Brown(?) was developing that was supposed to link hospitals and doctors surgeries and health centres, which was scrapped after allegedly spending up to £100 billion on it!dudeofdoomFull MemberIt’s strange thou the issue’s were fixablish and I’d have thought they would have wanted to do a Boris presentation.
If they’d swapped to the GApple API and put a proper agreement to not hang onto the data we’d have a product that other countries managed to ship and numpty boy wouldn’t be there blaming everyone else.
(I don’t doubt that our home built would be better at distance but Apple would be doing this with a compromise on battery life.)
Course the the big elephant in the room is whether any of it really works well enough. (App/public uptake)
Perhaps they needed games designers to make it something that appealed to the public as well as being useful to ensure uptake.
Like the football competition perhaps 🙂
sl2000Full MemberI don’t doubt that our home built would be better at distance but Apple would be doing this with a compromise on battery life.
It’s not a battery life issue – it’s privacy again. The UK app is sharing the phone model, then using a table of phone model against Bluetooth power to work out the distance better. Apple/Google’s API won’t share the phone model since it’s a personal identifier. This looks to be solve-able: phones could be classified into one of a few power bands (eg high, medium, low) and that could be shared without loss of privacy.
JamzeFull MemberHancock is openly blaming Apple, completely ignoring the fact that this app is designed to be platform-agnostic, able to function across a whole range of makes and models on iOS and Android! ****wit! 🤬
Yes, there’s both a Google and Apple version of the app. But the Apple version seems to be the one that everyone (not just us) is struggling to get working because of the way the OS is designed. Don’t think he’s saying Apple are causing issues on Android??
Singapore is now looking to develop a dedicated wearable device and roll it out to the population after all the issues they’ve had getting this working on smartphones.
SandwichFull MemberMeanwhile I’m using an app from Guys and ICL that allows you to report how well you feel day to day and has good privacy controls.
The problems come when developers go mad with data harvesting for things that they do not need (eg a battery charging app that wants access to my photos and address book, that’s a hard no and uninstall straight away).
We need an app that does the bare minimum with good privacy controls but those commissioning the work can’t stop themselves trying to see all that we get up to. The struggle with iOS is due to this compulsion to go where Apple say you must not to maintain privacy, the designers/developers need to “Think Different”.
kcrFree MemberApple have built their OS the way it is for various reasons, including security and privacy. Complaining that they are at fault and should change things to accommodate a specific app is a bit back to front.
It might just be the case that a mobile phone app is simply not a good solution for contact tracing, given the technical limitations of the operating systems. Compare the UK response to the pandemic with Kerala (a state of 35 million people) who relied on old school testing, tracing and quarantine:
Our “world beating” approach might give us a phone app without contact tracing by the end of the year…
JamzeFull MemberAn interesting counterpoint…
Just a reminder that in April Google and Apple together decided what covid19 proximity apps could exist, and how governments could use them.
A vital global public health policy decided with zero democratic accountability.
Privacy matters. Unaccountable power matters more.
— Tom Loosemore (@tomskitomski) June 21, 2020
ajajFree Memberhas good privacy controls.
If that’s the Zoe app then it’s a bit vague on the privacy controls; they talk about anonymous data but the information they collect is so detailed that it can’t be, they are vague on who gets the anonymous version, they are vague on whether things like IP address are considered personal or not. They don’t say what constitutes vital interests. The NHS gets the personal data anyway, for no good reason. They talk about contracts with Google and Amazon but don’t say if the liability is 2p or 2 billion pounds (it’s going to be closer to 2p). And they neglect to mention all the public agencies that are entitled to a copy of the data under the Investigatory Powers Act.
kelvinFull MemberJust a reminder that in April Google and Apple together decided what covid19 proximity apps could exist, and how governments could use them.
Bullshit. They only allow one app per a region, because otherwise the apps compete against each other, and none of them get the % uptake required to be useful. It is then down to the elected state government who supplies that app, and whether is uses the baked in APIs or not. If you haven’t got an app in your region, or it doesn’t work in the way you’d like, that’s is 100% down to your elected state government.
CougarFull MemberAn interesting counterpoint…
… from someone who clearly either doesn’t know what an API is or doesn’t care.
How much truth is there in the rumour that the developers were pals of Cummings?
Some were Faculty employees I believe.
JamzeFull Member… from someone who clearly either doesn’t know what an API is or doesn’t care.
But the key functionality is Apple/Google’s. You’re mandated to use their distributed design and contact detection algorithm (given any other approach won’t work), and it comes with restrictions.
With hindsight, we should have just done a Germany, roll it out and any contact detection issues are then down to A/G. Had a bit of early feedback from someone using their app, says it doesn’t seem to be doing anything at all.
kimbersFull MemberIs anyone surprised that Hancock is lying about this?
he’s made stuff up on the hoof from the very start of the crisis
the daft thing is, the more the government so obviously lie to us, the less likely people will be to trust them & use the app!
kelvinFull MemberYou’re mandated to use their distributed design and contact detection algorithm (given any other approach won’t work)
You’re not mandated to use anything. As is proven by all the contact tracing apps around the world that don’t use their APIs, and are in use, and working, and are in the Google and Apple app stores. The only restriction is that you or I can’t release our own apps via their stores, and confuse the public. One app, chosen by the authorities, per a region, that’s the restriction. There are working apps in other countries based on open source code, and bespoke code… it’s not just Google/Apple API based apps out there.
JamzeFull MemberAs is proven by all the contact tracing apps around the world that don’t use their APIs, and are in use, and working, and are in the Google and Apple app stores.
Technically I don’t see how that is possible on Apple unless they have the apps foregrounded 24/7 or jailbreak their iPhones.
Just get the bloomin’ thing released. You’ve lost any public confidence anyway. Surely some contact detection is better than none at all?
dudeofdoomFull MemberI think the window on public confidence on the app closed a while ago, the whole Cummings affair probably didn’t help with trust.
I’ve just recently seen the FB posts doing the rounds about the Covid app installed on your phone,( it’s the permission screen) wowsers 🙁
kelvinFull MemberTechnically I don’t see how that is possible on Apple unless they have the apps foregrounded 24/7 or jailbreak their iPhones.
It doesn’t work very well. So using the APIs makes sense. TraceTogether shows the limitations (and predated the APIs)… ‘our’ people claimed they’d found a better way than TT used to ensure/replicate keepalive… shock horror… they were bullshitting or just plain wrong… ours was worse.
I’ve just recently seen the FB posts doing the rounds about the Covid app installed on your phone,( it’s the permission screen) wowsers
2am WhatsApp message from the head at my other half’s school about the permissions… 😑
CougarFull MemberBut the key functionality is Apple/Google’s. You’re mandated to use their distributed design and contact detection algorithm (given any other approach won’t work), and it comes with restrictions.
You’re not “mandated” to do anything of the sort, rather that a framework has been created to allow applications to function in a secure manner. Plenty of countries are trying to do their own thing including the UK up until a U-turn a couple of days ago when they realised that (“told you so”) it didn’t work.
There are restrictions for what applications can and cannot do and this is a Good Thing. Do you really want the UK Government – a government who, lest we forget, introduced the Investigatory Powers Act, wanted to make secure encryption illegal and has a tame Big Data and AI company in its back pocket – to have direct access to the hardware on every mobile device in the country? Cos I sure as shit don’t.
CougarFull MemberIn case we had all conveniently forgotten about the Investigatory Powers Act or “Snooper’s Charter” incidentally, as a reminder here’s a list of all the companies that can demand your Internet records without requiring a warrant:
Metropolitan Police Service
City of London Police
Police forces maintained under section 2 of the Police Act 1996
Police Service of Scotland
Police Service of Northern Ireland
British Transport Police
Ministry of Defence Police
Royal Navy Police
Royal Military Police
Royal Air Force Police
Security Service
Secret Intelligence Service
GCHQ
Ministry of Defence
Department of Health
Home Office
Ministry of Justice
National Crime Agency
HM Revenue & Customs
Department for Transport
Department for Work and Pensions
NHS trusts and foundation trusts in England that provide ambulance services
NHS National Services Scotland
Competition and Markets Authority
Criminal Cases Review Commission
Department for Communities (Northern Ireland)
Department for the Economy (Northern Ireland)
Department of Justice (Northern Ireland)
Financial Conduct Authority
Fire and rescue authorities under the Fire and Rescue Services Act 2004
Food Standards Agency
Food Standards Scotland
Gambling Commission
Gangmasters and Labour Abuse Authority
Health and Safety Executive
Independent Police Complaints Commission
Information Commissioner
NHS Business Services Authority
Northern Ireland Ambulance Service Health and Social Care Trust
Northern Ireland Fire and Rescue Service Board
Health & Social Care Business Services Organisation (Northern Ireland)
Office of Communications
Police Ombudsman for Northern Ireland
Police Investigations and Review Commissioner
Scottish Ambulance Service Board
Scottish Criminal Cases Review Commission
Serious Fraud Office
Welsh Ambulance Services National Health Service TrustStill want to trust them with your tracking data?
JamzeFull MemberDo you really want the UK Government…to have direct access to the hardware on every mobile device in the country? Cos I sure as shit don’t.
Has the swear filter broken??
We are talking about allowing an app to scan and connect to a BLE peripheral in the background.
andytherocketeerFull MemberWith hindsight, we should have just done a Germany, roll it out and any contact detection issues are then down to A/G. Had a bit of early feedback from someone using their app, says it doesn’t seem to be doing anything at all.
Probably because, the German app was released about 4 days ago, only about 10% of the population downloaded it in those first couple of days, German new infection numbers are now pretty low (with the localised exception of 1 outbreak in an abattoir etc.). So assuming a perfectly even distribution of app downloads, that’d make something like 200 people in a population of 80 million trigger the alert in the last week, of which pretty much all 200 people would have already been showing symptoms and already be in bed within 24 hours of getting the app, where most/all of the matches will more than likely be people in the same household or place of work.
So not seeming to be doing anything at all, statistically seems about right. Unlike NHS IoS app where not seeming to be doing much at all was completely predicted and publicised by app software devs.
kcalFull MemberYep on the Covid applications auto installing Facebook posts – *really* _ this gets reposted without critical thinking, a quick check and yep its not really true – but after Cummings, the government and publicity about NHSX, Track and Trace, Google and Apple, its a plausible post.
CougarFull MemberHas the swear filter broken??
That isn’t and has never been in the filter.
We are talking about allowing an app to scan and connect to a BLE peripheral in the background.
You might be.
CountZeroFull MemberSomeone on Twitter has suggested we might have a beer ordering app in place before we have a C19 track and trace app. Sounds good to me… 😉
CougarFull MemberAsk yourself this.
Google and Apple – combined the OS creators of pretty much every still-supported smartphone handset in the world – have collaborated to come up with a method of providing anonymised data in a consistent and standardised way to the designated app of every state leader in the world.
In the middle of a global emergency where speed of response is critical, the UK instead chose to commission AI experts to write an app from scratch using unproven ideas which bypassed G/A’s distributed model and would collate all this data (allegedly) anonymously in a centralised database that they’d have 100% unrestricted access to.
Why?
kcalFull Memberbeer ordering app – was already in place for likes of Wetherspoons I think.
mate was involved in a remote ordering app previously, which had only partial uptake – if it can be adapted for pubs and cafes, could be on to a winner!ajajFree MemberWhy?
There is always “Never attribute to malice that which is adequately explained by stupidity”.
GCHQ has had mobile phone tracking and tracing since before the Omagh bombings, so it’s not that. Maybe they don’t want people to know how good it is.
CougarFull MemberThere is always “Never attribute to malice that which is adequately explained by stupidity”.
Hanlon’s Razor. I’d be more inclined to believe that if they didn’t have previous. Lots and lots and lots of previous.
kcrFree MemberIn the middle of a global emergency where speed of response is critical, the UK instead chose to commission AI experts to write an app from scratch using unproven ideas which bypassed G/A’s distributed model and would collate all this data (allegedly) anonymously in a centralised database that they’d have 100% unrestricted access to.
Why?
Because the government are crap at doing their job. See also every other aspect of managing the Coronovirus pandemic.
NorthwindFull MemberJamze
SubscriberYou’re mandated to use their distributed design and contact detection algorithm (given any other approach won’t work),
That’s not being “mandated”, that’s just how you work with an existing product. It’s like saying your car manufacturer mandates what fuels you can use. Or, if you’re Matt Hancock, filling your car with pepsi max then blaming Toyota for not being more supportive of your cola-car plans.
CougarFull MemberOr, if you’re Matt Hancock, filling your car with pepsi max then blaming Toyota for not being more supportive of your cola-car plans.
That gave me a laugh, thanks for that.
ayjaydoubleyouFull MemberOnly 68 StopCovid users have declared that they’ve been diagnosed COVID-19-positive in the app. The server received 205 social interactions from those 68 cases.
In the end, StopCovid has only sent 14 notifications over the past three weeks.
So about 3 interactions per infected person, seems reasonable. Also:
On the server again, each user is associated with a risk score. If it goes above a certain threshold, the user receives a notification. The app then recommends you get tested and follow official instructions.
I guess not every fleeting bluetooth contact is deemed infectable.
But… 1.9million using the app out of 60 million odd, how many contacts with app-less people were there?
dudeofdoomFull MemberBut… 1.9million using the app out of 60 million odd, how many contacts with app-less people were there?
It’s one of the elephants in the room with ‘relying’ On an app.
I Think you’ve got to do a ‘tie in’ say a weekly/ Daily lottery to entice people to want to use it, give prizes out.
Perhaps have more than one app, say a football one all doing the tracking but providing something useful,a version For each of the Countries favourite interests.
Having a working app is one thing but having one out there in large enough numbers to be effective….now that’s hard.
kimbersFull MemberTransmission is much, much higher in enclosed spaces, you just need to get the people in these places
Halfcock was halfway there when he said you’d have to leave contact details at door of pub.
Instead only allow entry if you’ve downloaded app, same for churches, public transport, hospitals, offices, swingers clubs & other high risk venues
The topic ‘The contact-tracing app, accuracy?’ is closed to new replies.