[OSM-talk] definition of the main highway-tag

Martin Koppenhoefer dieterdreist at gmail.com
Fri Jul 31 14:28:06 BST 2009


2009/7/31 Greg Troxel <gdt at ir.bbn.com>:
> I object to the notion that there should be a different relationship
> between residential/unclassified in urban vs rural areas.  We already
> have too much of that, and I think it's a sign our definitions are off
> base.  There's no clear boundary, and we have to translate this to
> garmin, etc., use in Free nav programs, and render, so people doing
> things differently based on where they are or what they're used to seems
> like trouble.  That said, I see the trouble with the secondary/tertiary
> definition (will send separately about that).

Well, I just see it as a hierarchical line:
residential
unclassified
tert
sec
prim
trunk
motorway

it's simple as that, and I don't see any problem.


> To me these are both
>  real streets that you can drive on
>  roads you would probably only use to get to places near them
> and the only difference is that residential means it's mostly bordered
> by residences.

nah, not all streets where someone lives nearby are residential
streets. They are just then residential streets, if they are small and
used only/mainly by residents. Big streets with external traffic are
never residential streets, even if people live there.

> Arguably the whole notion of highway=residential is
> somewhat broken, since residences nearby should be landuse=residential
> polygons, but it does affect the feel of the road and it runs pretty
> deep in osm, so I won't really object.

It's an easy way to speed up routing calculation and to improve the results.

> Perhaps we need a specific highway=alley tag to say "this is a road you
> can drive on, but it's definitely narrow/inferior and you don't want to
> go there unless you have to in order to get somewhere".

we already have this. Highway=service, service=alley.

cheers,
Martin




More information about the talk mailing list