Have stumbled via OSMI about this monster in Portugal, which isn’t really manageable anymore. It tends to freeze weeker systems and creates timeouts, e.g. no chance for the mapki history. And osm.org throws an “application error”:
http://tools.geofabrik.de/osmi/?view=areas&lon=-8.10308&lat=41.51186&zoom=12&opacity=1.00
http://www.openstreetmap.org/browse/relation/5389229
name = Carvalhais das Serras de Fafe
landuse = forest
dominant_taxon = Quercus robur
leaf_cycle = deciduous
leaf_type = broadleaved
type = multipolygon
My only chance to get it fully loaded into old system was to download a reduced version without metatags via overpass (4.7MB):
overpass-api.de/api/interpreter?data=relation(5389229);%20(._;>>;);%20out;
This is NOT a case of shreddered MPitis. Most (or all?) rings are 1 closed way.
It is also NOT one big forest with tons of inners.
Instead it consists of only a few major areas but plus countless little ‘satellite’ forests: over 900 OUTERS!
Cannot check each and everyone of those easily, but from looks and a few examples assume that most are just little confetti forests. Tiny, independant patches of forest, often for no more than a few trees.
It is my understanding that there are no superrelations possible for multipolygons, but is that thing really necessary all COMPLETE in 1 relation? Just to have the same NAME for all of it? Considering that name doesn’t even seem so important, at any rate a google search doesn’t even find any hits for the exact expression (“Carvalhais das Serras de Fafe”)
Or could this 1500-member monster be split up into a couple dozens smaller relations, making it manageable again?
Additionally, is it really justified that even the tiniest of all those forest patches must be in a relation at all, claiming them all “Quercus robur” trees?
