[OSM-talk] Routing algorithms and areas

Robert (Jamie) Munro rjmunro at arjam.net
Tue Jul 17 12:00:28 BST 2007

Hash: SHA1

Iván Sánchez Ortega wrote:
> Hi there,
> I'm having some doubts about the current route-planning algorithms, and about
> how do they deal with areas.
> For example, I might want to park my car in a parking, and then switch to a
> footpath leading to my destination (e.g. a building).
> Another example: I'm walking, and the optimal route crosses a square (or
> plaza), or a park (which I can walk out of the designed footpaths).
> Problem is, the road leading to the parking is not connected to the footpath,
> so they do not make a network "per se".
> So, routing algorithm boys: how do you plan to deal with this kind of stuff?

When I mark a car park as an area, I am always careful to put little
highway=service roads linking nodes in the car park to nodes in the
street. I haven't done that with footpaths, but in general I haven't
surveyed footpaths, only roads. There's no reason not to do that with

The only interesting issue is that routing algorithms need to figure out
that if there is a town square with a footpath on each side, the cost
for a pedestrian is the straight line across the middle, not the length
of the outside, but this is a minor issue.

Robert (Jamie) Munro
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the talk mailing list