|
I use osm-inspector not only to verify my own work but to correct that of other contributors if they are within my areas of interest. I far as I understand the Multipolygon instructions touching-inner-rings are allowed in OSM in a multipolygon contruction: See http://wiki.openstreetmap.org/wiki/Relation:multipolygon#Touching_inner_rings However the osm-inspector displays touching-inner-rings with the comment Are these then errors and if not what is the osmi trying to tell me? |
|
Touching inner rings are a violation of the OGC "Simple Features" specification and therefore some programs that process OSM data might have a problem with that, and discard the polygon that has touching inner rings. The standard OSM rendering process that is based on an osm2pgsql import of OSM data into a PostGIS database has no problems with touching inner rings. You can switch off the "touching inner rings" layer in OSMI if you want. Thank-you Frederik for the prompt reply. I know I can switch it off but that is not what disturbs me. I understand now that the errormsg is not an error at all and it begs the question Why is it flaged?. Isn't this errormsg redundant, i.e. what purpose does it serve? If the errormsg "touching-inner-rings" is irrelevant, how many other OSMI-*errormsg* are also irrelevant?
(29 Apr '12, 14:03)
dcp
2
The OSM Inspector doesn't only show errors. It shows different "views" of the OSM data. It is your job to interpret those views and decide what to fix and how to fix it. The field called "errormsg" is a bit of a misnomer, just think of it as "description" or so.
(30 Apr '12, 10:52)
Jochen Topf
|
|
In my opinion, dcp correctly concludes that a "touching inner rings" case is not an error (I would avoid "redundant", not enough strong). Even refering to "some programs ... might have a problem ..." is confusing if we accept the often used OSM statement "we are not tagging for renderers...". Most programs will today easily handle the referd "touching inner rings" case (and much more complicated that apears in OSM area classes). I like your point: "we are not tagging for the renderers". I would still like to know how many other errormsg flags are also irrelevant as I have not only been falsely "correcting for OSMI" but I have also contacted others contributers citing OSMI and erroneously pointing out their errors. This is not encouraging!
(03 May '12, 06:43)
dcp
|
