Also a "fabulous" piece in the Times about the NHS app-ocalypse ttiled "Coronavirus tracing can stop second peak, insists Patrick Vallance" (see https://www.thetimes.co.uk/article/coronavirus-tracing-can-stop-second-peak-insists-patrick-vallance-2s2jkhpcf).
It illustrates a decentralised app (Google / Apple) style uploading ids to a centralised database. Firstly, the Google / Apple api uplaods encryption keys (souceL https://en.wikipedia.org/wiki/Google_/_Apple_contact_tracing_project) for the last 14 days, not ids.
The article goes on to say: The app can be “woken up” from tens of metres away by Bluetooth from Google-made Android phones and by iPhones that are using the tracing app in the foreground. The Bluetooth used in most devices is class 2 and has a range of about 10m. In addition Android phones will only be able to operate in this way if the app is running and in the foreground (same true for iPhones) or the app is running a foreground service.
It also says, "The government could resort to asking iPhone users to actively turn the app on before leaving the house." That won't get round the problems because the app will be pushed quickly into the background as the user almost certainly does other things. It also stops working if the iPhone is locked.
So it's a poor article which shows a lack of understanding. However, it does manage to describe some of the pitfalls of the UK's approach to the contact tracing app. Some of the problems are intrinsic to the chosen architecture and are not bugs which can be fixed. There will of course be some of them too.
Originally Posted by: Brian Gaze