DP-3T / documents

Decentralized Privacy-Preserving Proximity Tracing -- Documents
2.25k stars 180 forks source link

Feature Request: Human rememberable/configurable numbers as option #122

Open jjtmp opened 4 years ago

jjtmp commented 4 years ago

Some days ago I published a similar concept, see http://jbechtel.de/site/sonstiges/covid19/CCfC/ (It just consists of a tiny description and a brief "FAQ".)

CCfC is not oriented on individuals carrying numbers but rather events/shops/locations. It also has somewhat more moderated approach, so it assumes everyone, including shop owners etc. can get a bunch of IDs at local authorities desk and then chooses a new ID every new day or every 3 hours or so. This ID will be displayed at the entry door, so even people without technical skills may write it on some piece of paper.

In order to make it possible to join physical concept (CCfC) and electronical concept (DP-3T) I suggest to explicitely to add the option to make the EphIDs freely configurable (so allow even for short ones).

mrseeker commented 4 years ago

The question here is: If you mark the store as being "infected", how many people would avoid the store? And how do store owners prevent themselves from getting "marked"?

jjtmp commented 4 years ago

That's a short question, but difficult to answer.

Let me first try to guess what you are asking for, making some kind of story of it:

  1. People in general don't want to get infected (as we are waiting for medical immunization) - this fact is often implicit in the tracing discussion, as pointed out by some comment in #118 .
  2. If you (or your store) have an infection, other people don't like you, even if you wear a mask.
  3. For the store owner it is an economical desaster when customers change to competitors (because of 2.).
  4. The technology of EphIDs/CCfC IDs (or whatever you may call them) makes easily available to the public where an infection was
  5. Maybe (as you write it here in issue #122) you want to indicate that CCfC is even more plain open as there's no cuckoo filter and it's more tangible

I agree with these aspects 1.-5.. It's a nasty situation.

But there are inherent and additional measures to prevent bad rumours:

There may be situations where competitors try to find infected-but-not-yet-reported people and send them to stores they want to harm. But this "game" can also be done with EphIDs or with these other tracking techniques (EDIT: centralized PEPP-PT version? - but I don't know the details). This "game", however, doesn't work if you get rid of the virus (or live with it) without contact tracing apps. (Again refer to #118 )

Did I guess your question right?

Remark (a): If the administration manages them well - so gives a bunch of 20-50 completely different numbers to the owner. We must trust administration in this point. If we don't want to trust, we could also individually create some form of MS Windows GUIDs but they are longer and not so practical. This of course is a choice which every city would have to make on its own.

Remark (b): which hopefully don't get misused.

jjtmp commented 4 years ago

Ah, just to mention: I implicitely have stated that DP³T and CCfC are ~equal, this was just because I read #177 and #163 before - that are techniques which are also yielding to reach "places", not persons.