[LINK] Australia’s Rollout of Covid-19 Tracing App Is Marred by Secrecy and Bugs

Glen Turner gdt at gdt.id.au
Tue Jun 2 18:42:04 AEST 2020


> There is no “Goldilocks” zone in the middle of these two extremes.

Bluetooth signal strength and time the signal are seen are both
continuums. So why choose fixed extremes in the first place?  What's
wrong with the sort of careful mathematical tuning we apply to other
real-world models -- this problem isn't a million miles from
cryptographic traffic analysis, where of course you incorporate other
data (such as the person's recollection) and apply different parameters
for analysis in one situation than another (such as on a train versus
walking the street).

Moreover, analysis need not be exact and need not happen only once. For
example, if you've identified a "super spreader" then re-analyse the
data and ask all contacts to be tested, whatever the signal strength or
elapsed time acquired.

I'm not defending the app, but if criticisms are going to set up
strawmen, then they should not present their conclusions as applying to
the real world.

-glen

-- 
 Glen Turner





More information about the Link mailing list