[OSM-dev] Thoughts on an enhanced GPX api

[Cartinus]

On Wednesday 29 July 2009 15:20:26 Ævar Arnfjörð Bjarmason wrote:
> That's because JOSM has a feature to automatically connect points that
> are near each other, but that's as far as it goes since it can't get
> any track information from the server.

On Wednesday 29 July 2009 15:20:59 Martin Koppenhoefer wrote:
> but that's not relyable, because AFAIK JOSM just "guesses", which
> points belong to one track, it doesn't really know about.

The api returns gps points ordered by the time they were logged. So there is 
not much "guessing" needed and it is almost always right. AFAIK there are 
only two reasons where things can go wrong.

1) Two tracks logged at the same time. Even in a large metropolis this very 
seldom happens.

2)  JOSM behaves as if there is only a single track. But this is fixed by 
dropping all segments over a certain distance. So only if two traces 
(recorded right after each other) intersect with the edge of your bounding 
box close to each other will you see an "invalid" segment. Since you are 
probably more interested in the centre of you boundingbox .......


So while these new ideas give you access more data about the GPS tracks, 
saying that now you just have a useless cloud of points in busy areas is way 
off reality.

-- 
m.v.g.,
Cartinus