OpenStreetMap logo OpenStreetMap

Changeset When Comment
122875457 over 3 years ago

The area is also a sawtooth shape, which is pending to be added. It will be easier to draw this as a `=platform`, rather than `barrier=`, or `area:highway=footway` and `area:highway=service`.

For other countries, eg in Berlin, there are even `highway=bus_stop` (no `=platform`) on a `highway=platform` + `public_transport=platform` (where the `platform` goes) . It's not a settled topic, including `=bus_stop` vs `=stop_positions` That's why there are discussions.

I review every changeset, and apply the same standards to myself. I don't delete anyone else's work lightly. I always try to improve on them. One reason being, there is one user here who keeps deleting everything to redraw them, often at a lower quality and missing tags; which others have complained before to no avail, and nothing can be done because technically it's not "wrong". It's not good practice.

You removed the area with reason being it looks bad. That's why I remind you osm.wiki/Good_practice#Don't_map_for_the_renderer By the same token, don't delete objects you don't understand either. osm.wiki/Good_practice#Don't_remove_tags_that_you_don't_understand

122875457 over 3 years ago

"people adding relations to them arbitrarily" is because there weren't any `=bus_stop` points until you added them now, thank you. " had become an eyesore after new buildings are built over it." is a renderer issue.

122875457 over 3 years ago

`=platform_edge` would represent the entire usable length, similar to `bus_bay=` on the road where they exists. (There are not even standards for bus stops per se) You considered boarding, but not disembarking. Other countries would use multiple doors, and even articulated buses.
You can look at HZMB to why `highway=platform` and `=platform_edge` may be needed. They are separated into platforms A alphabets onwards, then each position gets A1 numbers onwards. Each route can board at any slots on their area. Is it totally accurate to claim there are that many bus stops, for reasons of many stopping positions alone? As for door position, what about accessing the luggage bay? There will also be discrepancies in the stopping position every time. That's why both door position and `=platform_edge` is useful. Of course, here it is indeed one bus stop for every position.
Other countries use `highway=platform` on roadside stops as well. Which I don't agree with entirely, but it does show the variations.

122875457 over 3 years ago

1. They are not mutually exclusive. As I said, you can use `area:highway=` for the exact walkable areas. You should not delete something simply because it is inaccurate. The platform represents the entire area. There are many `=multipolygon` that can removed by this suggestion of yours simply because they need to have more `inner` added.
2. Again, this is a matter of iterative refinement. As I remembered, there were no individual `highway=bus_stop` + `public_transport=platform` added at first. Therefore I used the `highway=platform` first. You are definitely welcomed to add individual points to replace the `highway=platform` as members. `=platform_edge` only augments them.

122875457 over 3 years ago

For comparison, a railway platform would be occupied by other structures as well. However, the meaning of `railway=platform` is the entire area for it, including everything on it. Walkable areas should use `area:highway=footway`.

122875457 over 3 years ago

You are not allowed to directly copy from Wikia and Wikipedia. Please at least use KMB's official website only. (Although there are still grey areas with database rights, so you should rely on it alone)

122875457 over 3 years ago

1. OSM works by iterative refinement. That's why you can change it to a `type=multipolygon` and add `inner` members for the column.
2. This is a renderer issue. Don't Map For Renderer.
3. `highway=platform` is not the same as for `highway=bus_stop`.

122875457 over 3 years ago

1. Please don't delete the platform area. They mean different things, and roads can also be added as areas.
2. Please don't mention irrelevant sources. Wikia has an imcompativle licence . This may infringe copyright.

122567153 over 3 years ago

4. Forgot to say this section is not the same as the western side. The middle is not spaced. Gaps are only on the slip roads.

122567153 over 3 years ago

1. It's about keeping the history.
2. You should not use the quality/standard of the road cross-section as the defining criteria. Function in the network should be considered.
3. We do not consider only heavy vehicles, although this is a factor. As an extreme example, a road allowing only cars is still a `=motorway` if it is such. Wing Tak St is a main access for Riveria Garden, and only access to Pavilla Bay and Tsuen Wan Sports Center. Wing Shun St to the south is certainly not only used by heavy vehicles.
4. Not sure what is your answer . You should split the part in between, and keep the 2 parts as `covered=yes`.

122706266 over 3 years ago

These are separately ones. It was already drawn. Please use `descroption:zh=`, not brackets.

122567153 over 3 years ago

For `covered=yes`, `bridge=yes` does not denote the extent of overhead. `man_made=bridge` area needs to be drawn, which is more effort than splitting `covered=yes`. The opposition to tagging roads under a bridge is mainly on wrongly using `tunnel=yes`. Contrary to documentation, there are debates. osm.wiki/Talk:Key:covered#Discourage_"covered"_for_linear_roads_under_bridges

122567153 over 3 years ago

Also, please try to move, or use Replace Geometry in JOSM, instead of deleting then redrawing objects.

122567153 over 3 years ago

Actually you should not use standard as the criteria for classification. Function in the road hierarchy is used.
Wing Shun St to the south is `=tertiary`. While Ma Tau Pa Rd onwards is `=secondary`, Wing Tak St is the main access from Tsuen Tsing Interchange. Therefore any reclassification will have to be considered as a whole in the district's network.

122254313 over 3 years ago

I don't get what you see as a problem when `addr:housenumber=` is used on the whole estate at the same time. It's certainly not used for a single detached house only. It's an address concept.

122254313 over 3 years ago

Ignore the example wrongly using `addr:place=` for land lot.
Here is a smaller list.
way/302152975
way/228008980
way/525654594
way/616823524

122254313 over 3 years ago

This is not an etymological discussion. `addr:place=` needs to be a village with its own house-number, in lieu of house-number associated with. Housing estates don't have this house-number assigned inside. Quite the opposite, in larger developments with their own streets (eg Meifoo, Whampoa, Taikoo Shing), they are assigned house-numbers on those streets.
Here is a list of housing estates with village house-number currently mapped. https://overpass-turbo.eu/s/1jqH
There are many other with `addr:place=` but not yet tagged with `addr:housenumber=`. You can query for them, and compare other sources.

122254313 over 3 years ago

`addr:parentstreet=` + `addr:street=` is inaccurate, and doesn't solve the issue. An `addr:building=` would be more suitable for a building itself.

122254313 over 3 years ago

This is not the case. `addr:housename=` is same as `addr:housenumber=`, which in HK will be applied to the entire housing estate. Even if you find it misleading, `addr:place=` is for villages with `addr:housenumber=`. In fact, there are housing estates in villages using village house-number, meaning a conflict.

122467400 over 3 years ago

Secondarily, points on opposite directions are aligned to each other.