relation/9267547 still broken.

2021 .. relation still broken.

Hi,

Some more thoughts on this valley stuff.

Possibly leave the natural=valley alone as the base of the valley and add a new feature for the valley area/extent/? That would them fit in with making the new feature a simple area without a center bit that I don't think gets used elsewhere in OSM? So natural=valley_area ?

HI,

multipolygons are for areas ... not un-closed ways. So Kanimbla Valley (12110843) is an error.

For the moment I would map that part of the Coxes from top to bottom as the valley.

Nortons Ck looks to be part of the Megalong Valley ...

Hi,
the tag natural=valley should be on a way, not a relation.
See natural=valley

Good luck.
I walked the Gross just before the fires, would like to go back but the eastern access points are still closed.

No. Do large scale if you want.. but try not to reduce the detail.
The Great Western Highway could be represented with 2 nodes, one at the start and another at the finish and a simple straight line between the two nodes. It would be there in its simplest form while the least amount of noise. But it would be useless for navigation ... OSM should contain the detail, but not the noise ... and that is the fuzziness of the real world.

Good luck.

Hi,
According to the wiki natural=valley should be a simple way along the valley floor ... not a closed area of the valley peaks with the valley floor...
This does throw errors on OSM Inspector.

At present most other State Forest in NSW are tagged with landuse, the trees are not mapped within them, though some have said they will map them.. but there is a lot to do!
There is one sawmill operated by 'forestry NSW' (or what ever it is called now) tagged landuse=industrial IIRC. And the Cumberland State Forest - mainly a PR place with some research facilities .. not certain what that is tagged now.

Note: most renders take landuse=forest and natural=wood to be the same thing and show them as the same thing. So for most there is no point to the argument... but me - I do render them differently.

Yep, there is a lot to do and ways on top of ways don't help.

More on "overfitted". There is a tool in JOSM to 'simplify ways'. It modifies a way so that the result is within x meters of the original and reduces the number of nodes used. The default value is 3 meters, I have found that can le3ad to a result that does not resemble the original, where as 1.5 meters works. I usually use it on the result of another tool to map water bodies called scanaerial. I'll try running it on the way your complaining of and see what it does... later.

Hi,
I was unaware that you changed the State Forest from landuse=forest to other tags. As the primary function of State Forests is to produce timber the tag landuse=forest is best. While State Forests have secondary functions for recreation and nature conservation the tagging should reflect the primary function of the feature. Note that landuse does not necessarily imply a land cover. It is the entirety of the State Forest that gets the landuse tag, not simply the area of the trees.

If you want to tag the area of the trees seperatly then use the tag nature=wood ... note that the key 'natural' applies to both natural and unnatural things!

Ways 888019420 and 888019432 had no track information, nor were they part of the formal boundaries of the State Forest.

Rendering - see osm.wiki/Rendering

IF your quick you can see what attracted my attention at https://tools.geofabrik.de/osmi/?view=areas&lon=145.45605&lat=-30.87017&zoom=6 It will get up dated some time tomorrow .. say 3 pm~

However I have have re-included the school tree area. And I have roughly updated the tree area so it doe not cover the non tree areas .. until the next updated imagery comes along.

There would be more benefit in detailing the long straight bits of this tree area. Mater of time and inclination.

Excellent!

I exclude 'state forests' from tree areas because from time to time they will be harvested, so from time to time they will have no trees. I also render them differently - and having both things on the same area confuses things.

Hi,
What do you mean by 'overtraced'? If the meaning 'too detailed' ... detail can be removed by a program, but it cannot be added. So more detail is better. Look at the straight lines of the tree area south of Winmalee as an example of poor detail.

Spiting land cover boundaries because they cross some administration boundary (eg a school) is not required and adds some confusion..

This changset broke the relation for the trees and that has lead to a partial reversion in order to restore the trees.

As stated in the second comment .. "I'll retag them as leisure=nature_reserve", and I have done that.

I note that one of them is named a 'park', whatever that means locally. Those that have other names carry as part of their name 'reserve'.

The choices are ?

Not to map them in OSM at all ... despite knowing they are public lands.

Map them, but without the full technically correct detail that is not shown on our source.

------------------------------------
I take the second option, rather like mapping a building=yes where the kind of building is unknown. Any other ideas?

Also conflates boundary with NP boundary.

natural=wood is for a tree area, Some bits of these may not have trees. So I would not use this.

I'll retag them as leisure=nature_reserve, not certain of their protection status. They are clearly shown on the LPI/DCS base map. I would like them identified as being 'public land' so that anyone trying to deny use of it is know to be wrong.

Hi,

There is an error on one building.

Firstly Relation: 11991963 "type"="multipolygon" "building"="house"
has only one member and that relation can be deleted.

Second Relation: 11991964 "type"="multipolygon" "roof:colour"="#F2EEE7" "building"="house"

referred to the above relation .. but with no role. Instead the outer way for this building should be referred to with the role 'outer'.

Hi,
Relation: 11863112 "landuse"="grass" "type"="multipolygon"

has only one member and that member has the same tags... suggest you delete it.

Note I find these errors using https://tools.geofabrik.de/osmi/?view=areas&lon=151.75568&lat=-32.94248&zoom=12