[OSM-dev] coastline-osm2pgsql, once agai

SandorS sandors39 at gmail.com
Wed Jun 27 09:30:17 UTC 2018


Sven:
>>
Thus it has been decided a long time ago, that land polygons are rendered from
shapefiles proven to be mostly correct (e.g. all continents do exist) and
up-to-date stuff produced by osm2pgsql.
If you like to render your own map you can easily produse matching data from
the same planet file.
The coastline processing toolchain is available here:
https://github.com/osmcode/osmcoastlin
>>    
Martin:
>>
you can also get precompiled (fixed, splitted and not, 4326 and 900913)
water polygons, land polygons and coastline shapefiles here:
http://openstreetmapdata.com/
>>

Of course.
Actually, all examples of the erroneous small holes in the outer land polygons in the paper referenced from my mail are taken/based on the polygons referenced by Martin. In turn, these polygons are created by the procedures referenced by Sven. This really large number of issues is equally reflected in the mentioned water polygons as well, being splitted or not. 
But, please, do not miss the essential message related to the large/huge number of logical errors caused by the uncoordinated, unsynchronised editing and processing of the different object classes or feature layers. Mappers uploading a new hole/island in a river never know whether this should be done to the Global Ocean (coastline) as well. Likewise, when uploading a river section it is never checked whether parts of the existing river line run outside the river (creating imaginary islands) or not, the same when uploading forests over water areas and so on. These errors are passing most/all today publicly available error detectors, inspectors or logical preparation procedures. They are present in most of the publicly accessible OSM maps. Fortunately, the OSM source data has the potential for control and reparation of these errors and that was my main focus and intention by the “coastline-osm2pgsql” example in the mentioned paper.
Sandor


Sent from Mail for Windows 10

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20180627/d615d182/attachment.html>


More information about the dev mailing list