Add ability to view the number of currently visible exposure notification IDs

[ljl-covid]

No description provided.

[chris42]

Isn't something like that completely against the philosophy of microg, which tries to reduce data gathering as much as possible?

[chris42]

Sorry for not putting that correct: Exposing this information or make it available to use out of context. Data protection is 2-ways, so if I want to reduce my data footprint, I would of course do the same for other peoples data. So having a function, that gives the opportunity to log all IDs and use them out of context would hurt that?!

[chris42]

Thank you for lecturing me in this weird way. Your are correct, we do not decide this, hence I wrote my opinion and asked. It is up to Marvin to see if it fits microg and if he wants to put effort in it.

[chris42]

You figure, that this is an open space, in which what you post can be questioned and criticised? If you cannot cope with that, you should probably get in contact with the developers in a more private manner.

[mar-v-in]

I do agree that such functionality could be useful. However there are two main reasons I don't think it makes sense to have it in microG.

• It's not in any way necessary for the usage of microG and can be implemented outside of microG without any issue.
• Live data is not available inside microG. We only scan every 3 minutes and will probably change this to 5 minutes at some point. While this is very close to live, it could confuse users to not see a live number increase as soon as they jump in a full tram or whatever.

If you need data for.debugging purposes, you can either use external apps or use dumpsys to gather insights from the service.

[ljl-covid]

Okay, that makes sense. But please don't make the checking interval too sparse... some (most?) of these apps are already being way overly cautious with what counts as a "contact", both in terms of distance, which is inevitably guesstimated using signal strength, and time, which I believe needs to be at least 15 minutes, at least here. 15 minutes is a pretty long time, considering the distance criteria has been shown to result in two people kissing not counting as less than 2m apart simply because they had their phones in their back pockets... if to that, you add a check only every 5 minutes, even the 15 minute threshold becomes harder to actually meet.

If you're just uncomfortable to leave it at something that isn't what Google uses, please consider making it user-configurable: I really don't care if I waste a little more battery in exchange for better likelihood of catching contacts.

[mar-v-in]

I do agree that sparse scanning is going to lower the sensitivity of the whole approach.

However the original Google implementation only checks every 5 minutes, which is why many apps using it decided to only consider those that have 5, 10 or more minutes of contact, as the provided data accuracy is just not high enough if you want to reach a certain specificity

[ljl-covid]

FWIW, I've had the chance to look at a non-microG phone, and it does provide some kind of export (although I didn't check what data exactly would be exported):

So while I agree it's not necessary to use microG, the Google implementation actually seems to provide it to some extent. To be honest I'm more just reporting this simply because I'm surprised to find it.