More information regarding possible infection

[r-r-liu]

I see from the sample screen shot of the exposure warning that the app doesn't give very much information about when the possible infection occurred. I keep a manual record of where and when I go somewhere. If the app were to warn me of a possible infection, I would contact the health department of the my home canton and would like to be as helpful as possible. The time at which the critical exposure is supposed to have occurred would be helpful. I can only imagine two reasons for not supplying it: (1) To protect the identity of the person putative infector, or (2) Because the app is computing the exposure risk by cumulating exposures to different people thoughout the day.

[UBaggeler]

As you guessed, Apple and Google are withholding this information for privacy reasons. This makes it harder to infer which person that you met was tested positive.

[r-r-liu]

Let me try to get this straight in my mind. The app seems to have nothing to do with the broadcasting of beacons and the logging of encounters, is that correct? Otherwise, I would have thought that the app could keep track of the time that it received beacons, and wouldn't need the time that the now infected user broadcasted it. The only way to prevent the app from doing that is to have the GAEN api doing the logging and give the app no access to it. But then the GAEN would have to do major parts of the exposure risk calculation, not only parts pertaining to the proximity, but also with those pertaining to duration and even accumulated duration (if experts should decide, e.g., that it's possible for two infected people to infect a single person, each for approx. 7.5 min.). That, in turn, would seem to preclude different countries doing it differently.

I took a cursory look at the exposure risk calculation of the German app, just because I had the impression it was described in more detail that the Swiss calculation, and I had the impression that it was done differently. But I am probably wrong if it's the GAEN that is actually doing the calculation.

[r-r-liu]

@UBaggeler: I reread the document on the calculation of the exposure score and had the distinct impression that the buckets for the three attenuation ranges return an estimate of the total exposure time over all encounters in the given range, regardless the covid code with which an encounter was upload and the approximate time of the encounter. Is this true?

Although, if I am correct, that still doesn't necessarily mean that Apple isn't doing so in order to thwart attempts at guessing the infected person't identity, there might be some epidemiological considerations involved, too. In a typical situation tailored to the app, commuters on a train, I am not sure whether a 15 min. community with the same infected passenger sitting opposite isn't roughly equivalent to a similar situation where one infected passenger sits there for five, and is immediately replaced by another infected one, then he/she five minutes later by yet another infected one.

[r-r-liu]

I have found a more explicit description of the exposure score calculation (https://github.com/admin-ch/PT-System-Documents/blob/master/SwissCovid-ExposureScore.pdf) that explicitly states that my impression is correct.