I have found the landuse polygons to be a definite hassle to map around,
in urban areas. Some mappers in my area go house by house, parcel by
parcel, creating hugely detailed multipolygons that also happen to be
wrong far too often. Trying to map around this mess is unpleasant.
The current landuse tagging leads to complex multipolygons which are
fragile, in part because not all mappers understand them. Layered land
use (similar to level) might be more readily understood:
The big area:
landuse=park
stacking=1
The small area:
landuse=beach
stacking=2
---------------------------------------
I would prefer something much more tied to an already mapped boundary.
As you mention: parcel boundary polygons are generally seamless. But
parcels are naturally tied to zoning (e.g. /permitted land use/) as
opposed to actual (/mapped/) land use. Thus a parcel could
simultaneously be:
zoning=residential
And:
landuse=commercial
Zoning data is often available as part of official parcel boundary data
(see santa cruz
<http://wiki.openstreetmap.org/wiki/Santa_Cruz_County,_California>). It
is pretty rare to have an (important) land use difference within a
parcel. But an example is a college campus which leases land at the
edge for a shopping mall.
If we could default a larger admin_level polygon to a landuse (e.g.
/residential/, /farm/) then it significantly reduces the mapping
clutter. At that point you just mark the exceptions. But it seems some
tool help would be needed.
For areas without parcel data, streets are a possible edge. Imagine
specifying that landuse is /industrial/ in the polygon formed by streets
A,B,C and D. Mappers could adjust those streets and without the hassles
of land use polygons. Again, too help would be critical.
_______________________________________________
Tagging mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/tagging
