Sorry for the delay: I have had other things to do. I do not care for your putting words in my mouth as this was not at all what I had said. There is a difference between not having and not wanting a way to represent a singularity of trees, a problem which can arise from a lack of understanding in how to map natural=wood. Now, I admit, there I could have put in a better wood outline. However, your workflow, by deletion rather than omission of natural=wood, is extremely problematic. A suggestion is for sparsely wooded areas, one may enclose nearby lone trees. Yet wooded areas of unspecified density may also enclose lone trees.

There are feature types which should and those which should not be represented. However, if wooded areas which happen to belong to a loneliness of trees is among the features which shall not have a representation, then one forfeits all spatial analysis of wooded areas by lone trees. However, most wooded area is not necessarily by lone trees.

If one wishes to render lone trees as dots on the map, then one can just as easily check whether multiple natural=tree nodes are enclosed by the same natural=wood area. Your method does not work in reverse; by omitting data, one does not get data back when one needs it, and your rendering is imposed on the user.

I advise you to reconsider your argument, which seems to be based on groups of trees, and to try not to delete any natural=wood in the future when it is in fact part of a wood-covered area as per the definition in the first link you provided. If you had read the article for which you have provided me that link, then you would have seen that you are wrong: a single tree node may always be enclosed by natural=wood. If I had been tasked with identifying groups of trees, then I would have chosen another tag entirely (perhaps group=yes). I appreciate your efforts here but do not wish to likewise subvert any of your changesets. If you have a suggestion for the project, you should take it up with the wiki instead.

Furthermore, I suggest you expand on the abbreviation "OSMI" or include in your profile(s) the meaning of each changeset code. I have also observed similar changeset comments under the name "basics". It is important to include more descriptive changeset comments so that others know exactly what you are editing.

Wissen Sie eigentlich dass ein Baum (natural=tree) unterschiedlich ist von einem Holz (natural=wood)? Also es gehören beide Features hin. Sie brauchen das eine nicht zu löschen.

Dear Mueschel,

I am keen on removing them.

However, I couldn't find it in the Overpass or any of the ways in the changeset. I was looking for description=logical or logical=yes or meta=logical_path.

I would regret having introduced those tags as they don't really elicit much meaning to me.

I apologize for any inconvenience.

I have been keeping an eye on local edits so as to fix stuff once I catch a break. The link has been showing correct routing, and it appears that the issue has been avoided. Thank you for taking time to fix it, and for understanding. 🐈

Hello!

Those paths are just crossings, and though not visually appealing on the homepage, removing them will break the routing service in a few hours like so:
osm.org/directions?engine=fossgis_osrm_foot&route=34.41445%2C-119.86589%3B34.41377%2C-119.86519

When buildings are overlapping like this, it is only a matter of time before someone catches it. osm.org/way/486178691/history

How could the iD editor be improved so as to discourage unintentional changes from a landuse to an overlapping building?

a) By clarifying overlap in the walk-through.
b) By disallowing overlapped buildings.
c) By showing a confirmation dialog.
d) By rendering buildings as 3D.
e) By enlarging or adding color to the inspector's back/forward buttons.

Something else? Thanks in advance.

http://overpass-turbo.eu/s/ZgH

Are these WiFi network names?

`ref=13850` would work, too.

You removed the highway=motorway tag! :p You know how I knew? Because the highway disappeared from the map.

Can someone please add Coastal Dispensary headquarters? website=https://coastaldispensary.com/location/santa-barbara-delivery/

What do you think: Marsh trail should have `bicycle=yes` or `bicycle=designated`[?](osm.wiki/Tag:bicycle=designated?uselang=en)

I thought this was Carlos house

Is there a website for The Dome Home?

Hi dekatherm!

I should have time coming up soon to fix a limited number of these buildings by hand.

Ideally, we want to conflate data from street-level photogrammetry as well as from satellite imagery.

To this we should answer: is a building that is outlined by its roof boundary considered incorrectly mapped?

Only aerial or street-level photogrammetry can reveal the facades of a building, so any roof overhang could be represented by a building tag.

But by definition, a building is not a building:roof. So, I would like to tag the overhang of a roof as a connected area, though I have not seen this been done.

Otherwise, it should be sufficient to say by how much the roof overhangs with a building tag, though this may not fit the accepted tagging scheme. Walls, too, should be mapped as well as windows.

Details like this are not only fundamental to OSM as a project but would help define the metrics by which open source tools can find and replace malformed/missing building parts.

We should observe that as time progresses, hand-drawn or pre-existing OSM building data becomes an excellent training set for machine learning because of ground-truths.

I see what u did here. U started mapping a building in iD as a way instead of an area, and couldn't find the building feature type. U can undo this by adding the tag area=yes (or in this particular case, building=yes) to ur closed way. Then, u will have unlocked the search for all area feature types (and have hidden all the way/line feature types) Buildings are usually mapped according to the wiki.osm.org u will find that buildings aren't represented in osm using ways is all.

In this particular region u can get a sharper image by selecting Esri World Imagery (Clarity) Beta. If u scroll down to the bottom, u can boost the sharpness which can help aid the eyes. Then, if after outlining the building boundary u want to reposition it to your favorite reference imagery, just select the feature and press 'M' to move it to where it aligns with the Bing aerial imagery (as it is usually the default). In some regions, it doesn't make a difference.

The most important buildings to map imo are those that you wouldn't expect AI to catch on an aerial imagery.

U chose a difficult feature type to map. When u are done with a building, u may optionally square the area of the building by selecting the feature and pressing semicolon (;) to make it look neater.

The comment speaks for itself in this case.

Thank you.

Sorry for late reply.

Part of problem was recent urban developments and lack of imagery which btw Digital Globe now reflects though not yet hosted by OSM. So I didn't know exactly where to place the courts until now. I just updated it.

The other part of the problem was not knowing exactly how to tag a court, as a park, garden or grass, for example.

I changed most of the grass to garden because indeed it was unwalkable. You are right. I convolute or "split an area into two" when it is entirely subdivided by a concrete or asphalt path in an urban area. I'm reading the wiki as I go along, fixing the tags of nodes and their ways, relations. Please point me in the right direction if you see a mistake!

Most of the inspiration for convolution comes from the patches of green I see in UCSB campus maps where the areas serve to validate the ways by giving them edges. It seems like a good idea for such a large mapping project where mistakes can easily be made.

What happened to Ramona Court?