Hi there, would `note=level: 2 through 6` be better represented as `level=2-6`?

way/1235498283 ...`=secondary_link`? Why?

Are you sure about the `ele=`? Out-of-copyright maps say otherwise.

Feels like a descriptive name; are you sure?

By the way, do you also know if node/7134763063 node/7134763668 and node/7134763667 exist irl? (Didn't notice them at first as they don't render prominently on carto)

Is it appropriate to use `layer` this way? I quote layer=*

> Ways in buildings (or similar structures like multi-level parking lots, shopping centers, airports, railway stations, some multi-level bridges and roads...) should be mostly described with level=* instead of layer. level=* is used for the physical vertical arrangement of levels (floors, walkways) inside structures. Use level:ref=* for the human readable signage used to identify physical levels (eg elevator buttons, multistory parkade levels).

I take that to mean you should not use `layer=n+1` when there are no `layer=n`.

This is not a criticism or a complaint, just a question purely out of curiosity: how did know this alignment is correct / more correct than the old one?

Underground, you can't use GPS or aerial imagery, and you probably can't even dead reckon because there are no points of reference in a dark tunnel.

1. > The government websites include "Yau Ma Tei" https://www.info.gov.hk/tb_chest/tb-chi/contents/c614_1July2007.htm https://www.dh.gov.hk/english/tele/tele_chc/tele_chc_shcf.html https://www.dh.gov.hk/english/tele/tele_chc/tele_chc_shcm.html

I think we can simply believe in the website. Quoting you from note/3209599, "I don't think you need a survey for this. If the (official!) website contains wrong data then it's entirely their problem and we have no obligation to help them clean it up."

2. Understandable. But then, `=clinic` is an obviously wrong tagging, and with how few people are seemingly interested in this topic, I don't think you can ever find an obvious and uncontested replacement without more attention. Maybe post on the forum?

1. Yes, that's what I linked, I have expressed my views there am I am still awaiting your response after 24 days.

2. I think at least we can have a consensus that it is not `amenity=clinic` or `healthcare=clinic` due to the duplication problem.

See changeset/175801448 ; latest situation reflected

> The "contraflow bus lane" can be simplified as "this road will only lead into the bus terminus".

No it can not. E.g. if a private cars want to travel from Chai Wan Road to node/316735012 , then it would need to use Hong Man Street - Cheung Lee Street - Chui Hang Street - Lee Chung Street then back to Hong Man Street. Adding `oneway=no` would mislead data consumers into thinking they can route straight on.

> Normally the "this is bus terminus, do not enter" signage is placed at the proper entrance, but for convenience and reminder purposes, it's placed earlier.

I don't think this particular situation is like that. I feel like it's to prevent wasting buses time when enter Greenwood Terrace BT, so they made a short section of contraflow bus lane.

> It's like you wouldn't need a "no left-turn" OSM restriction when the left side is oneway coming out.

Yes, and similarly you wouldn't need to "only right turn" restriction from Hong Man Street to Cheung Lee Street because there was already `oneway=yes`, but you removed it.

> Ideally routers should learn to recognize "impossible dead ends" and exclude them from general navigation.

Yes, but as I said above, a route can end during that short section. And even if routers work, it is still not an excuse to add incorrect information.

> I prefer things to be simple.

I don't even know why you made this changeset. You aren't creating it; if you want things to be simple then just leave it as is, it won't bother you. It doesn't make sense to actively replace a more precise tagging with a less precise one, let alone a wrong one.

Isn't it a `short_name`? Afaik `loc_name` basically means a nickname that only people from a certain small area (usually the area around the object) knows about.

1. Are you sure they are simply "Student Health Service Centre" and "Special Assessment Centre" with no "Chai Wan" prefix? See changeset discussion of changeset/174721133

2. way/777216821 should probably be `landuse=healthcare` or `=institutional` and not `amenity=clinic`.

> `motor_vehicle=no` was added in the past, not me.

Didn't notice that, sorry for the wrong accusation then. However we can still discuss on the correct tagging.

> Old style is `vehicle`=*, new style is `motor_vehicle`=*. This is mostly iD editor changes

I have no idea what you are talking about. Currently iD's GUI lists options for All, Foot, Motor Vehicles, Bicycles, and Horses. Do you mean in the past the "Motor Vehicle" field was just "Vehicle"? But then bicycles are vehicles so that does not make sense.

> "vehicle" and "motor vehicle" in practice are largely equivalent (at least in HK).

No? Bicycles are vehicles but not motor vehicles. As the name implies, motor vehicles are a subset of vehicles. See motor_vehicle=* and vehicle=* .

In Hong Kong, I think we can simply use the definitions given by Cap 374 Road Traffic Ordinance:
"vehicle(車輛) means any vehicle whether or not mechanically propelled which is constructed or adapted for use on roads but does not include a vehicle of the North-west Railway or a tram"
"motor vehicle(汽車) means any mechanically propelled vehicle"

In that case, `vehicle=no` would correspond to traffic sign 113 (white circle with red border), while `motor_vehicle=no` would correspond to traffic sign 114 (white circle with red border and diagonal slash with a black motorcycle and car symbol). (go to Cap 374G Schedule 1 to look at the traffic signs)

...what is this trying to do? Is it not a contraflow bus lane?

Do they use village name addressing? Should it be `addr:place` instead?

Are you sure it's `motor_vehicle=no` and not `vehicle=no`? Most bus termini are `vehicle=no`.

Yes that is my guess too, which I why I changed some of them to `leisure=park` when it is "obviously" the case, the ones I deleted were unbelievable or added by users with a history of adding non-existent features.

Actually they aren't; there are no gate/signage at all indicating a slope maintenance path, the bottom of the staircase is directly connected to a random sidewalk.

Yes, indeed some routers (e.g. OSRM) do assume `bicycle=no` for `highway=trunk` (see https://github.com/Project-OSRM/osrm-backend/issues/1996), but I am not sure if `bicycle=yes` can override it.

Now even though it is said that we should not "map for the router", `bicycle=yes` on roads where cycling is legal is not really incorrect or fictitious, and it increases the chance for data consumers to interpret data correctly (e.g. would they really look at `not:bicycle=no`?), so I don't see a reason to not do it.