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.

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`.

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

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

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

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.

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.

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

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.

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

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.

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

As I said before, please don't move lines upstream beyond the physical separation. `change:lanes=` is already used here.

Please don't remove the address.

Please don't change `=pedestrian` to `=footway` when it's a emergency or maintenance vehicular access.

Please don't add bracketed descriptor labels in `name=`. That's for proper names only. Use `description:zh=`.

HI, can your team stop removing `highway=footway` + `area=yes`, or at least change them to `area:highway=footway`? For example this one is a wide open sea in front of the mall with a brand statue, so it's natural someone wants to draw it. Can you share what's the validation issue.

Please don't extend turning lanes to sections without physical separation.

Please check the lanes. This is required on top of `change:lanes=`.

Please don't use Taiwan name of Sukiya. HK has no such branding.