You should not make assumptions about others.

I fixed the issue in the most appropriate manner, since we don't have a well establish alternative relation type for collections.

If you disagree with changes, you can just revert them and make the changes you see fit. Arguing in changesets is not a productive use of our time.

When your polygons don't get flagged as old-style or broken by osmium, they won't get on my radar.

http://area.jochentopf.com/

Because it wasn't a valid multipolygon, it has touching out rings. A collection of rings like this cannot be modelled as a multipolygon.

The individual rings were tagged with those from the relation to fix the invalid multipolygon.

Why did you not inspect the changes in the object history?

In the context of the area project every relation with only type=multipolygon is old-style.

http://area.jochentopf.com/

You're wrong, removing the relation _is_ a fix for old-style multipolygons. The object won't be detected as an old-style multipolygon again.

Instead of arguing in this changeset, you should have inspected the history yourself and taken the actions you see fit.

I did, and the change looked sensible, it should have deleted the relation instead of just removing the tags.

With your revert you risk the duplicate information being removed from the relation again.

It had no tags after the duplicate tags were removed in the changeset before mine:

relation/9349785/history

Disconnecting the node is sufficient to fix this.

Does iD not allow you to do this?

It's fixed with changeset/70519490.

What are the IDs of the nodes in question?

All settlement (woonplaats) boundaries are relations, no exceptions.

Osmose should be updated to not complain about single member boundary relations, those are perfectly valid. And very much preferred for consistency with other more typical boundaries with multiple members for the outer ring.

The tagging of both the member ways and the relations are correct. We've used this scheme for the administrative boundaries for several years, and we're not about to change because you or Osmose think differently.

Please spend your time on more worthwhile efforts. The administrative boundaries are well maintained.

Don't delete boundary relations. Only having a single member is not a problem.

The tags were indeed copied as-is from the relation. Please go ahead and fix the tagging.

Various types of broken polygons, see:

http://area.jochentopf.com/

Issues include wrong role for multipolygon members, touching outer rings, duplicate section in polygon, self-intersections, etc.

For these relations it was mostly wrong roles, specifically inner members that were outside the relation.

Don't create new (old-style) relations for the new wetland members, add the wetlands as inner members to the existing relation:

relation/9512149

See the relevant documentation:

* osm.wiki/Relation:multipolygon
* https://josm.openstreetmap.de/wiki/Help/Action/CreateMultipolygon
* https://josm.openstreetmap.de/wiki/Help/Action/UpdateMultipolygon

Your relations have no tags, nor do most of their outer ways.

Looks like your import is broken.

As you can see in the history of the relation, the landuse tag was removed leaving the multipolygon without tags describing the feature.

Those look like disconnected ways resulting from the Replace geometry action.

Because those ways were not modified when the boundary way was replaced, the validator doesn't check them.

I don't consider these nodes a problem as they belong to different ways and should not be nodes shared between those ways. In my JOSM layer they aren't exact duplicates, their high precision coordinates are different.

If those nodes do bother you, see my previous comment.

I don't find any issues in the datasets I use (which are limited to the administrative boundaries).

Other features should not be connected the boundaries, so I advise you to move the other features away from the boundaries instead of merging duplicate nodes.

Please do not "correct" administrative borders in and of The Netherlands. We use authoritative datasets to maintain those borders.

Please do not "correct" administrative borders in and of The Netherlands. We use authoritative datasets to maintain those borders.