[NUUG kart] [Imports] kartverket imports to OpenStreetMap

[Christoph Hormann]

On Wednesday 16 October 2013, Gnonthgol wrote:
> >
> > Then it would in my opinion make sense to tag all code 3 as
> > waterway=river and code 2 as river or stream depending on manual
> > assessment.  No width tag.
>
> Having looked at the data this makes little sense, it is common to
> have a width 3 "river" flow into a width 2 "stream". But the data can
> be quite useful when mapping local areas anyway. It is hard to map a
> river or a stream with only a gps device. This goes in the "vector
> overlay" category instead of import.

If you think this information is completely unreliable you can of course 
ignore it all together.  But keep in mind the river/stream distinction 
is not an importance rating, therefore it is fully possible for a river 
to run into a stream.

> > One other thing i saw when looking at the wiki page: There seem to
> > be distinct data sets for river and lake polygons and there also
> > seems to be a data set for dams.  It would make sense in my opinion
> > to set water=reservoir for all lake polygons bordering to a dam and
> > water=lake for the rest.  This could be done automatically but of
> > course also by hand when merging the data.
>
> Even better, there are different way type for regulated lakes as
> opposed to natural lakes. There is no distinction in the polygons
> though (SOSI have a topolagical data model like TopoJSON or OSM). But
> this is possible to make.

Note existing import tools already allow identifying ways shared between 
touching polygons.  If there is additional useful information in the 
topological data sets that is lost in the conversion to polygons/line 
strings this would of course be important to fix for any use of the 
data.

Greetings,

-- 
Christoph Hormann
http://www.imagico.de/