- From: Martin Thomson <martin.thomson@gmail.com>
- Date: Tue, 12 Aug 2014 21:22:29 -0700
- To: "Michael[tm] Smith" <mike@w3.org>
- Cc: Garvan Keeley <gkeeley@mozilla.com>, Ravikumar Dandu <rdandu@mozilla.com>, Prandstetter Josef <j.prandstetter@mysynergis.com>, Giridhar Mandyam <mandyam@quicinc.com>, public-geolocation@w3.org
On 12 August 2014 18:52, Michael[tm] Smith <mike@w3.org> wrote: > To be more clear, what I meant here is more precisely that the logic which > ends up being used to actually choose which sources (GPS, Wifi, whatever) > for determining the location is sometimes (often?) logic down in the > chipset -- through a location function of the chipset that even down that > level abstracts away which particular means of determining the location are > used (similar to the way the Geolocation API abstracts it out way higher up > at the Web API level) (Of course after that the actual means the device > uses for doing a location query is through some hardware on the device.) This is entirely true. A few years ago, while working on systems that did this, hybrid trilateration was what I was working on, taking measurement data from multiple sources to create a single value. This is now commonplace, so reporting that you have a mix of GPS, ground elevation model, radio timing and a bunch of other things is only useful in the narrowest use cases that really belong elsewhere. This is an old decision that the previous incarnation of this working group made, and they made the right choice here.
Received on Wednesday, 13 August 2014 04:22:57 UTC