position accuracy

[Jussi Kukkonen]

Henri Bergius wrote:
> On 8/8/07, Jussi Kukkonen <jhkukkon at cc.hut.fi> wrote:
>> First, I can't imagine where lat and lon would have (meaningfully)
>> different errors... This is not available from gpsd as an example. Maybe
>> this was a mistake and the errors were meant to be for horizontal and
>> vertical?
> 
> So: accuracy_2d and accuracy_3d?

Yes. Of course that is mainly theoretical as the API currently has no
way to query vertical position...

> Also, I'm not sure if meters are really the way to go here.
> 
> In Midgard I've been using just an integer that gives approximate info
> (1=very accurate, 9=maybe city-level accuracy).

Two (fairly weak) arguments against that:
1. gpsd does have "gps_fix_t.eph" and "gps_fix_t.epv" (for horizontal
and vertical uncertainty) available in meters. Those values change over
time as satellites come to view, so we could argue that our
position_error is not a momentary value but a constant description of
the uncertainty of the backend...

2. value in meters might be easier to use (no need to consult a header
file to see what value 6 means).

-jussi

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 307 bytes
Desc: OpenPGP digital signature
Url : http://lists.freedesktop.org/archives/geoclue/attachments/20070809/5f695bfb/attachment.pgp