Am Sat, 31 Mar 2018 14:05:11 +0200, schrieb Libor 'Poutnik' Støí¾:

Android applications and I suppose iOS ones as well
do not determine location on their own,
but ask the OS what the location is.

Hi Poutnik,

That's correct, I don't think that's necessarily the only correct answer
since an app can maintain its own database of whatever it wants to maintain
(e.g., some apps track in a tunnel where they use predictive algorithms).

In the case of the Google Map app, we all know that Google maintains a huge
*online* database of BSSID:location pairs for all access points that have
been reported to it by most (not mine) Android phones.

Everyone on this thread already knows that it's *possible* that Google
simply downloads that BSSID:SSID:location:dBM data to the portion of map
that is downloaded for offline use with the "OK MAPS" mechanism.

The main question is how to prove whether that BSSID:location information
is actually stored inside the *offline* Google Maps database which is
downloaded by the "OK MAPS" mechanism.

Does it track you only within ranges of supposed Wifis ?

I need to run tests, but the preliminary answer is "yes".
It seemed to work best in populated areas moreso than in remote areas.

If that proves out, that would be a hint that Google Maps is using a
previously stored local database of BSSID:location information, which, by
the way, may be the main reason that Google offline maps *expire* after a
month (actually 29 or 30 days, depending on the platform).

Does it even track you without any wifi in the range ?

Preliminary results are that "houses" need to be nearby.

Whether or not that means access points is unknown, but what else can it be
using since there is no GPS possible nor is there cellular data possible?

I think we all suspect that Google maps is downloading to the "OK MAPS"
localized database the access point information that most Android phones
(not mine) hand it every day, but the question is how to prove that is the
case.

In case of the former,

Does it track you only at known locations,
where the device has been before ?

Ah. This is an interesting question, where the tracking was certainly done
in an area that I frequent, since it was "in town", as I was leaving the
library (where I had been connected prior).

The Google Map app does ask if it can get location information even when
the user is not using the app, so, there "may" be something in your
question for us to track down further.

Thanks for thinking of that, which nobody else had mentioned (AFAIR).

Does it track you even at new locations
with wifis unknown to the device ?

That is a good question. I will have to use the tablet in a location where
it has never been before to ascertain that answer.

But, I think we all suspect it simply downloads the BSSID:location
information, which is the same information I queried a long time ago to
prove we could track someone's cellphone from our desktops at home.

You have to have a "key" which Google provides for free, and then you give
Google *two* SSIDs (the reason for two, as I recall, is a rudimentary
security mechanism), and then you give it a bogus signal strength, and if
things match up for Google's liking, it hands back the location.

Of course, Google's own map app is probably not restricted to two SSIDs
even as they would want as many as possible for their "triangulation" to
work reasonably accurately.

It seems to me not probable
to have pre-downloaded all wifi location database.
Perhaps, it may predownload the data for nearby area only.

That's is the conundrum.

How does the Google Map app manage to track you when the tablet has no GPS,
no cellular, and the WiFi doesn't have the credentials to connect to
anything but an access point which is wide open and just happens to be on
the road, but you'd pass by that access point fleetingly so there isn't
much time to connect to the net.

Most of us already suspect the "OK MAPS" mechanism downloads this
BSSID:location information - but the question is how to prove that is the
case.

I would rather believe the tablet does not behave
in the way you suppose it to behave.

That may very well be the case.
And I appreciate that sentiment of yours, which is reasonable.

I think we all know of a good mechanism that might be what it uses; we just
need to figure out how to prove that is the case.

Sometimes a particular HW is present,
but not available for the user rather for business reasons,
being either deactivated or just not used.

This might be the case, but I doubt it (but it's always a possibility).

It is cheaper to produce 1 version
and selectively disable optional features.

Yup. Apple does this all the time, e.g., with their modems where they have
an iOS device with a fast modem and another with a slow modem, so they
permanently slow down the fast modem just so that the user won't see the
speed difference in the two devices.

But this would be permanent, and hence not useful to the app.

Or, selectively disable it just for user applications,
but not for the OS.

This might be the case. I don't see any evidence of this though, in this
case.

E.g. Xiaomi MI A1 does have the FM radio chip,
( one can listen the radio via the service menu )
but it is not mentioned in manual
and there is no builtin nor downloadable FM application
that would work.

This is an interesting example. So it might be the case here, but I doubt
it since Occam's Razor tells us that the simplest solution is the most
obvious one that we all know already, but which we don't yet know how to
prove it's the case.

Q: How can we prove that the Google App "OK MAPS" mechanism downloads the
BSSID:location information of local access points?