Criticism whlie saving timestemp

[kreativmonkey]

In the comments of this article link are critics on saving the timestamp with the token. We notice this critic and would like to discuss if we can improve that issue.

[kreativmonkey]

The idea behind storing the timestamp is that we need an indicator when the TCN (temporary contact number) was sent and when a TCN was recorded. There is no indicator for this in the TCN itself, so we need to store this information locally on the device. This information never leaves the phone and is never reported to anyone else like the server.

In my opinion, an improvement is not to store the entire timestamp. We can store the time like a day or a part of a day, so you don't know in which second you give it up.

[Dionysusnu]

The sent-out TCNs get changed every [interval], correct? If so, you don't need to store the recorded-TCNs timestamp, the infected device just needs to send only those sent out at times at which it was able to infect others. You'd still need to store the sent-out TCNs timestamps, but I don't think you can really avoid it. After all, you need to store somewhere, when the contact happened.
If not, there's still very little point storing it to the exact second, the minimum necessary would be [interval].
[interval] = how long a TCN lasts before the app starts sending out a new one.

[haveyaseen]

It might not be necessary to store the contact times more accurately than to the day. In the end we want to be able to figure out the risk for infection. So if this risk is 0, that is if the user should no longer be able to infect others, those TCNs should get deleted. For both use cases we only need to know roughly when the contact happened.