On Fri, 23 Nov 2018 at 17:35, Eugene Alvin Villar <sea...@gmail.com> wrote: > > On Wed, Nov 14, 2018 at 5:30 PM SelfishSeahorse <selfishseaho...@gmail.com> > wrote: >> >> 1. 'inner' roles (and thus 'outer' roles too) are still needed in case a >> country has enclaves. > > > Even if a country has exclaves and/or has enclaves within it, you still don't > need to have "inner" and "outer" roles at all in order to make sense of the > (multi)polygon. They are there as a hint to fix hopelessly broken > multipolygon relations, but if such a relation is not broken, the "inner" and > "outer" roles are actually superfluous. For instance, the program osm2pgsql > actually has a check function named "check_inner_outer_roles"[1] to identify > relation member ways having the wrong roles in a type=multipolygon or > type=boundary relation. It is able to do this check precisely because it is > able to analyze a multipolygon relation and be able to infer the correct > roles for itself. > > We should be therefore able to repurpose the roles in a type=boundary > relation to store information about claimed, "de facto", and "de jure" > borders. > > [1] > https://github.com/openstreetmap/osm2pgsql/blob/93b73e5f5c3b20e80027ecf272f553d26f49f2e8/contrib/libosmium/osmium/area/detail/basic_assembler.hpp#L172
Thank you for this information and for correcting me! _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
