šŸ”Ž šŸ§‘ā€šŸ¤ā€šŸ§‘ The Coronavirus Contact Tracing App

I was in Copenhagen in December … at a conference… with about 3000 people from China and Asia… I came back and had a dry cough and felt a bit shit for about 4 weeks over christmas and new year…

giphy

1 Like

It’s what we’ve all been thinking, Bob, but none of us quite had the heart to say it. You complete and utter bastard.

4 Likes

When your phone appears to go to sleep, does it continue to receive emails, WhatsApp messages and the like? There’s your answer.

That’s no good. His head would sort of… bounce up and down above the water.

2 Likes

Ah, ok, hadn’t thunked about it like that. So are we saying that the track and trace app works off mobile data ? So what’s all the fuss about bluetooth or have I got the wrong end of the stick ? So the app is purely for recording if I have the symptoms or tested positive. The physical location of my phone is done by mobile data ? Is my GPS location that accurate that it can distinguish someone within 2m of me ? I’d be surprised.

Good point. If it’s Bluetooth and not data (Wi-Fi or mobile) then I wouldn’t know off the top of my head if it turns off when the phone isn’t active. But, given that the phone remains generally active even when you’re not actually using it, I’d think that Bluetooth is always active as well.

How much battery is this going to eat

People are going to quickly uninstall when they find the battery life halved

1 Like

I agree @CB-Saint, but I also think that people will soon work out ā€œhowā€ it works and if (as we are supposing) it’s via some kind of Bluetooth system, I reckon people will just turn off BT when they are out and about … hey presto, I wasn’t anywhere near you ?

It will also be interesting to understand what happens if you are told to self isolate … will there be a tracking system linked to your phone’s GPS which alerts someone if you go outside your house ?

I appreciate it all sounds a bit Big Brother but ultimately this boils down to the general public to use it correctly. If that’s what is required, why not just educate the public properly now and let them run their lives themselves ? The minority of idiots who go to Durdle Dor for the day and sit <2m from everyone else will not download and use this app correctly so aren’t we back at stage 1 ?

My understanding. [apologies for long ramble, but needed to get my thoughts on this down somewhere]

So there are two ways to build the app (the architecture) and two ways to handle privacy.

App Architecture

On the app architecture, the first distinction to make is between apps and the operating system.

The operating system (Android or iOS) is incredibly optimised to do very specific jobs and is written in a computer language that executes significantly faster than the language that apps are written in (general rule, some grey here).

If things happen quicker at the operating system level, many times quicker, then that task will use much less battery. For this reason, the operating system is the place where you want to put things that NEED to run all of the time. It’s the most efficient place for those tasks to be managed.

Secondly, most apps that aren’t interacting with a user sit idle until an event occurs that brings them to life, usually for a brief period or time, before sleeping again. Each one of these events, no matter how short-lived, will use a chunk of processing time and therefore battery. So the fewer events that your app responds to, the better the battery life.

You can, at least on Android, give apps permission to ā€œrun in the backgroundā€. This means that instead of responding to the odd event here and there, the app gets to see all (many more) events and appears to run all of the time. It actually doesn’t but the effect is that it appears to be running all the time. The consequence of this is that if that app acts on lots of the events its seeing, it will use lots of processing time and therefore consume lots of battery.

So if you were to write an app that needed to continuously find out if it was in the location of other phone users you have an architectural choice.

Firstly, you could write an app, give it permissions to run in the background and every ā€˜n’ milliseconds you say (probably over bluetooth)

ā€œam I near another user running the app?ā€
ā€œam I near another user running the app?ā€
ā€œam I near another user running the app?ā€
.
.
.
ā€œam I near another user running the app?ā€

You get the idea. This processing of continually asking for information, often in vain, is known as polling. Computers do it all the time.

This polling will eat battery, because apps are generally less efficient and accessing Bluetooth will need to physically broadcast and receive wireless stuff which again has a ā€˜cost’. On slow phones, the performance left for other apps may noticeably drop off.

Secondly, you could put all of this polling down at the operating system level - where things run fast and there are fewer layers of code to work through. This will still consume battery because Bluetooth broadcasting is battery-expensive.

The problem with this approach, if you’re a national government, is that the only people that can put the code into the operating system are Google and Apple - both sworn enemies.

BUT, in an amazing act of corporate selflessness, Google and Apple, very early on in the pandemic, announced that they would collaborate to provide these operating system services to app developers.

There was a catch, and that brings us on to privacy.

Privacy

So, a condition of Google and Apple’s largess was that they would not enable any personal or human identifying information to be collected during this process. This meant that they would not support sending information about who has potentially been infected, or their location at the time to a central database.

Why did Google and Apple act this way?

Perhaps because they would be sued out of existence (well, maybe), if any of that private information were to somehow get in the public domain or be used by a government for illicit purposes.

This meant that the privacy architecture that app developers would have to use would be different.

Here’s an excellent diagram that shows the difference between centralised and de-centralised approaches.

From this article.

It basically means that your phone itself does the working out who you’ve been in proximity to (nothing is sent to a central location) and it also sends the alerts.

Our government, for some reason, initially turned its back on the collaboration between Apple and Google, but when they realised that their approach would either be ignored due to privacy concerns or drain the battery due to poor app architecture, then they had little choice.

Interesting Dan but I still stand by my concern that ultimately it can be defeated ? Someone (assuming they have downloaded the app in the first place) can decide when the app is on or off to suit their own needs ? It boils down to honesty and integrity, both of which are demonstrably in short supply in the general ā€œpopulaieā€ …

You’re right. But the app isn’t the only measure that succeeds or fails by honesty.

If you are self-employed and just got back into work after a long lay-off and you are told to isolate for 14 days by the ā€˜human’ track and tracers. You’ve got a choice.

No app involved. Just a citizen weighing their interest against the nation’s.

On the app, I believe that having it running may be a condition of entry/travel in some countries. But ultimately, I think you’re right. It’s an honesty thing.

Make it a condition of going to a pub / restaurant - have a QR code on the entrance that you have to scan on the way in and out

That way you get these businesses open and I will definitely use it if I can go and get a meal or a pint

Yeah, especially as I am one of those and also thrown under the bus by Rushi. Fuck it, nothing is stopping me from working …

Yeah, not sure that would work with this Bluetooth polling model.

As @Saint-CD said, what is to stop someone turning off the app when it suits them.

Having the app on to enter the pub but turning it off as soon as you got inside would mean the last two people you were in proximity to were the bouncers.

IF it worked on GPS location and IF that information was sent to a central location for processing, then your approach would work.

But then I’d be unlikely to run such an app on my phone.

(I might be missing something in the way it works. Perhaps it can’t be turned off once you’ve signed up to it?)

They use it in Singapore in all public building including offices, shops etc - you have to scan in and out, so if you did turn it off they would still know where you had been. But by turning it off, you may end up having to quarantine even if you didn’t come close to an infected person on the premises

Singapore do back it up with a strict self isolation policy, with checks and severe punishment

If you refused to down load it then you would be restricted to what you could do

People still use Bluetooth?

As I say, that’s a different system.

The Google and Apple system works solely by storing the anonymised proximity of others and sharing your anonymised proximity with them.

It has no idea where you are.

I think I’d be happy with a system such as Singapore that I could choose to give my location to gain entry to somewhere - choose in the sense that I know I won’t get to go in if I choose not to share my location.

I don’t like the idea of a central system knowing all of my movements all of the time.

It’s excellent at > v5.0.

Took a while but it’s stable, low power and long range.

Is this about the quiz?

3 Likes

Absolutely. My very point exactly. :crazy_face: