I am either misunderstanding the context of this thread, or I am missing
an important technical point.
I get the potential value and complexity (in many dimensions including
privacy) of Covid-19 trackers.
If the thread is intended to encourage folks as individuals to help with
ongoing efforts to build such things, then okay, I can understand that.
(Although that is not what I thought I read.)
If the goal is for the IETF to do something, I am missing the technical
point. I do not see a protocol development or specification issue. The
task has lots of hard parts. Most in the application space and in the
data crunching spaces. (And probably other aspects that I am not
noticing, but that are also not protocol issues.) What is the IETF task
that is being asked for?
Yours,
Joel
On 4/15/2020 2:46 PM, Dirk-Willem van Gulik wrote:
On 15 Apr 2020, at 18:42, Keith Moore <moore@xxxxxxxxxxxxxxxxxxxx
<mailto:moore@xxxxxxxxxxxxxxxxxxxx>> wrote:
On 4/15/20 12:07 PM, Benoit Claise wrote:
Hi,
Which leads me to a question: what can this community (and
similar/adjacent ones) do productively together to help? What new
things are happening on the network from which we can learn and
quickly adapt/improve?
In my wish list, I would see this community helping with a COVID-19
contacts tracker:
- with clear specifications
- that respects the privacy concerns, for all parties
I don't think it's possible. Anything that can be used to trace
contacts for medical purposes can be used to trace contacts for
political purposes.
I would beg to differ, though by no means perfect or yet there, the EU
recommendations:
https://ec.europa.eu/info/files/recommendation-apps-contact-tracing_en
set out quite a 'hard' set of requirements; that by and large match the
manifest/expectations of the CCC, de Waag and similar more
activist/vigilant privacy groups:
https://www.ccc.de/en/updates/2020/contact-tracing-requirements
(I picked the DE one, as I could not find english version of the
substancially similar NL, FR, SE and DK versions) and designs such as de
DP3T design (with a few nits and warts) by and large meet those
requirements.
https://github.com/DP-3T/documents/blob/master/DP3T%20White%20Paper.pdf
<https://github.com/DP-3T/documents/blob/master/DP3T White Paper.pdf>
This is done by de-centralizing; and essentially constructing the
cryptography such that only 'on' the phone is it possible to reconstruct
'has there been a contact' and limiting the scope/purpose to exactly
that - have I been close. So no location, no tracking, no recording of
position, etc. And with sufficient means for an outside observer to
verify this.
The apple/google proposals are very similar - but are not as limited in
`time and place'; potentially more generic.
Now obviously - there is nothing stopping someone of using the very same
spec to accomplish something different; to spike the app, put hidden
code in it, etc, etc. But that is something that we have any way -
those that control the phone in your pocket can put a spy in your pocket.
Dw
