https://osmcha.org/changesets/144435963/

Please don't simply delete them as they represent a shop space, and may contain addresses. Change them to a plaint "point" to remove the feature.

Can you please stop creating fake `historic=` features? Other wrong examples don't justify more wrongs.

Ok what are you doing here except to break routing for the ability to cross the street there?

Also for those that are closer to `=footway` , you will need to reprovide crosswalks. Currently they can't be routed

Please don't disconnect `=pedestrian` from roads. They are pedestrianized vehicular roads.

Please don't delete the `landuse=` . It is for describing the site, or incompletely the lower part of building. It also contains the address that you removed together.

https://osmlab.github.io/osm-deep-history/#/node/11133588943

You accidentally moved Sai Sha Rd. Need to be careful when panning the map.

I'm reverting this fully as the `change:lanes=` info is removed. Realigning alongside.

As I said, please don't use the project title as a descriptor label on `name=` . That's for proper names only. Furthermore, this is the `landuse=residential` of the entire estate.

If it is signposted, there is a case to be made for using it in `name=` . But the most common name from addresses of the tenant seems to be "Facilities" . That's why I asked. Signs can also have problems. It's complicated.

Ok it's not. But it's the same minor mistake.

@Patrickov This is fixing a previous changeset. It's fine.

Are you sure?

Village names should not be added to the `landuse=residential` . Housing estates inside a village are separately drawn and named, but are still part of the village. `addr:place=` and `addr:village=` / `addr:hamlet=` (some village don't use village name address numbering) is enough.

On the contrary, this cause duplication on finding data centers. You still have to de-duplicate them. If there is no site drawn, only individual buildings, this becomes impossible. `telecom=data_center` is conceptually one facility.
Your example still doesn't qualify them. Does it matter if there's a separate building for the internet connection or control room? Is it "part" of the data center? Also for size, you want to count floor area, not number of individual structures.
It's very common flaw to specify such `building=` to be one-to-one with the facility. This won't do what you want, because people will still tag everything inside a `telecom=data_center` as `building=data_center` blindly. I recently discussed `building=fire_station` , and it may be used on all buildings, not only the main fire truck depot building with duty station.
It is a very common documentation shortcoming to require combination with `building=` . `man_made=works` was once such case, later corrected to be used on the entire site.
Problems with `building=data_center`:
https://taghistory.raifer.tech/?#***/building/data_center
1. Significant use come from mass modifications
2. Only ~350 in 2020
3. Relatively small number for a `building=`
4. Continued issue with British vs American English
I find it prudent to exercise caution on creating more `building=` .
Relatedly, it already won't work for `telecom=exchange` . `building=exchange` is meaningless. Creating `=*_exchange` for `=telephone_exchange` etc is unscalable.

Please don't add descriptor labels to `name=` . That's for proper names only. It should be akin to what it would be named after completion.

You misunderstood this. There is already `telecom=` on the `landuse=industrial` osm.org/way/854227561 of the site. That `industrial=` is for the `building=` , because I don't agree with `building=data_center` yet. This would cause duplicate `telecom=` features.

Please don't delete them directly, as they contain an address, and represents a shop space. Change it to a plain "point" , which removes the tags.