[Angstrom-devel] RFC: upgrade gpsd to 2.35

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Paul Sokolovsky schreef:
| Hello Matthias,
|
| Thursday, January 3, 2008, 3:11:52 PM, you wrote:
|
| []
|
|> Apart from the apparent dislike of the Author of gpsds networking
|> function, for which he never gave any good reason but "I don't like it",
|> an interesting read indeed. But as someone else already mentioned, gpsd
|> is the "standard" ATM.
|
|
|   I can second this. It doesn't take long thought to imagine how nice
| it may be to use GPS-enabled smartphone as a kind of standalone GPS
| for laptop, and doesn't take too long forums searching do see that
| that's what many users actually look for. For WinCE/Windows, one has to
| install commercial multiplexors to achieve that, fortunately, with Unix,
| it's standard feature, thanks to gpsd.

When people use the gpsd network mode, they usually just forward a 'raw'
NMEA stream over a socket. You don't need gpsd for that.

Andrews 'pty' app or 'cat <gps device or pipe> > /dev/rfcomm0' can do
that for you, 'nc' might even be sufficient. The rfcomm method is
probably the nicests since it's emulating a serial port.

What I think Iain doesn't like about gpsd (and its networking mode) is
that it invites people to shuffle NMEA sentences around, instead of
position data. This means that the cpu is going to decode the NMEA
stream N+1 times for N gps clients, 1 time for gpsd and 1 time for each
application.

In the long term we want to talk to GeoC1ue[1] to get our position (and
velocity vector for navigation), which could use gypsy, hostip or gsmd
as backend.

regards,

Koen

[1] http://www.freedesktop.org/wiki/Software/GeoClue

- --
koen at dominion.kabel.utwente.nl will go go away in december 2007, please
use k.kooi at student.utwente.nl instead.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFHfg/eMkyGM64RGpERAuJMAJ9PgWhXDXGVwoeKlv1KjH842AfPwACfQZmi
TW4vyF5bVbkDxgEHXNZu0Kc=
=gP36
-----END PGP SIGNATURE-----