[OSM-talk] Multipolygon fixing effort done

Jochen Topf jochen at remote.org
Tue Aug 29 14:16:51 UTC 2017


On Tue, Aug 29, 2017 at 03:54:44PM +0200, Christoph Hormann wrote:
> On Tuesday 29 August 2017, Jochen Topf wrote:
>  >
> > > Would the number of visible problems in the map due to dropping
> > > broken geometries now, after the fixing effort, be low enough so
> > > this change could be made to the main map to give mappers a better
> > > feedback about broken geometries?
> >
> > Osm2pgsql is switching to the libosmium MP code which is more strict
> > than the older code before. As far as I know the code is finished and
> > should be in the next osm2pgsql version.
> 
> I am aware of that but this does not answer my question if the fixing 
> effort has significantly reduced the visual impact rolling out this 
> change to the OSMF servers would have.  I would assume it has but the 
> OSMI does not really allow for a proper assessment here.

When the old-style-mp support was disabled on the main map, this left
about 50,000 broken multipolygons, in many cases clearly visible on the
map. I didn't see a single complaint about this. So, no, I don't think
being a bit more strict with broken MPs will be a problem with the map
or would be a reason to delay deployment of a new osm2pgsql version.
10,000 intersection problems in 300 million polygons is a rounding
error.

Note also that Mapbox has been using libosmium code for a while and they
are actively fixing large broken MPs every day to keep the map from
breaking in a bad way.

Jochen
-- 
Jochen Topf  jochen at remote.org  https://www.jochentopf.com/  +49-351-31778688



More information about the talk mailing list