Re: [gpsd-dev] [PATCH 7/9] Correctly define autonomous and A-GPS mode

[Sanjeev Gupta]

On Sat, Apr 18, 2015 at 1:34 AM, Greg Troxel <address@hidden> wrote:

I believe you.  But that's a good situation.  I have turned on a GPSr in
a car (so with some obstruction) and driven around 5km before getting a
fix.   Where I am has not only hills, but lots of trees, and things are
even worse in cities.

But that is not a typical use case for gpsd.

When I got my first mobile phone with a GPS (2005?), I was told by the manufacturer (Dopod, now HTC) that the reason for A-GPS was twofold:

1. user sees quick TTFF
2. which leads to lower power consumption

of which the second was the driving criterion: as soon as you exited the application, the GPS could be powered down *completely*, if we knew that we could pull data (over 12kbps GPRS) fast enough.

 As most of the recent work (this is anecdotal) seems to be improving gpsd for timing, and considering that even if gpsd is used by Android the A-GPS data will be fed in without our knowledge by the phone manufacturer, do we need to support it?

And each chipset (and each access method, USB, RS232, I2C) will need special magic anyway.

--
Sanjeev Gupta
+65 98551208     http://www.linkedin.com/in/ghane