[colug-432] precision of fixed platform browser geolocation

[Rob Funk]

On Friday, June 07, 2013 07:30:42 PM Richard Hornsby wrote:
> I was on my bank website at home on my Macbook, and asked it where the
> nearest ATM was.  Safari dutifully asked if I would like to allow the
> website, and Safari, to use my location.

You did ask for the *nearest* ATM. :-)

> How is Safari (google maps?) getting the location that precisely?

Safari isn't Google Maps. In fact Apple and Google have gotten into a 
battle over mapping, so they're no longer working together on mapping.

But as others mentioned, if there's a wifi access point nearby, that's a 
likely source of location data.

> Does
> google maps really have an internal map of private wifi access points
> and their locations?

Yes. So does Apple. And others.

A major use for it is to speed up GPS locating; GPS is slow, but can be 
made faster if you can limit the range of locations where you might be. If 
you could be anywhere on Earth, GPS could take a few minutes to figure out 
where you are; if you already know where you are within a mile, GPS is much 
faster to tell you more precisely where you are. So the location services 
use every possible source of location information they can to narrow down 
where you might be. One of those sources is the location of wifi access 
points. Since those don't change location very often, and you don't need to 
actually connect to them in order to get basic identifying information, 
they're useful for approximating location.

> Is this from the street view car that they
> "inadvertently" collected wireless data they sniffed?

The issue there was that street view cars were collecting wifi location 
data as intended, but *also* saving wifi packet data along with the access 
point metadata. The access point identifiers and locations were the point 
of the collection, and the packet data was the scandalous part. I can 
imagine that happening accidentally if someone used the wrong options on a 
wifi sniffer.

> Are they possibly
> sharing the access point information from their Google Fiber service to
> their maps/geolocation services and figuring it out that way?

Again, Google is not Apple. If you were using Chrome (or accessing a Google 
service), then that might be a possible source of information, except that 
Google Fiber only exists in a couple cities at this point.

> This can't simply be a browser thing alone - it is an OS level setting
> whether I want to allow geolocation services  to be enabled or not. 

Safari is very much a part of Mac OS X, and uses its native services.

> Anyone have insight into how this works?  If it is an OS protected call,
> is the OS making the request to somewhere for my location on behalf of
> and giving that to the browser?  If so, where is the information coming
> from?

It's probably making a request to an Apple server, sending a set of usable 
data that the server can turn into a location. (It probably wouldn't make 
sense for a database of wifi locations and cell towers to be in the OS.)

-- 
Rob Funk <rfunk at funknet.net>
http://funknet.net/rfunk