MC Navi released

[Al Johnson]

On Thursday 04 March 2010, Neil Jerram wrote:
> On 4 March 2010 11:11, Al Johnson <openmoko at mazikeen.demon.co.uk> wrote:
> > On Thursday 04 March 2010, Neil Jerram wrote:
> >> So that left me thinking:
> >>
> >> - given that gpsd already existed, and appears to Just Work on the FR,
> >> I wonder why FSO decided to write ogpsd + fso-gpsd from scratch?
> >
> > FSO is a collection of dbus protocols. gpsd at that time did not have a
> > dbus interface. freedesktop.org already has a dbus protocol called gypsy,
> > intended for devices such as ours.
> 
> I assume you also considered adding the gypsy dbus API to gpsd?  It is
> open source, after all.  It would be good to understand the reasons
> for not doing that, in order to evaluate whether they still apply
> today.

Check the smartphone-standards list archive. I was a spectator rather than a 
participant, and the arguments made at the time didn't seem unreasonable. They 
may have overlooked something though.

> > Adopting the gypsy protocol was an obvious choice at
> > the time, but the world has moved on.
> 
> What do you mean here?  Is gypsy now considered not to be a good API?

I just meant that if we look at things as they are now we may not come to the 
same conclusion. We now have expericence of ogpsd and fso-gpsd, not just a 
proposal. I think gypsy is mostly good (grr...integer timestamp...) but it 
doesn't seem to have gained much traction. Almost all apps are using the gpsd 
interface, so Geoclue on top of gpsd might be a better option. gpsd has 
changed too.