Reverted since I can't find a source that says this actually exists.

The reading of incline=up/down is based on the direction of the way (if the vertical direction changes then split the way where that happens?)

Maybe use one object for the red minibus (share_taxi) terminal and one object for the parking lot?

access=*

Oh. According to the wiki page it should probably be access=no (inaccessible to general public); access=permit isn't listed on that page and might imply that anyone could get a permit to use the path.

What are you referring to in the changeset comment? If the thing you've added as a hedge is actually a group of trees, then you should change it to natural=tree_row (natural=tree_row). Otherwise I can't see anything immediately wrong with what you've added, since all of it appears to exist from the aerial imagery, and nothing is terribly misaligned.

The convention is to use the name without "Station" or "station", since it is already clear from the symbol that it is a station. There is no hard rule but usage is fairly consistent across most regions as far as I'm aware. (The reason the bus stops use "Tai Wai Station" is that the bus stops are named after the station itself.) I will be undoing this change.

Are you sure that the slip road no longer exists? (you also forgot to delete the second half; see changeset/64449695 – you might want to use JOSM instead of iD if you're comfortable with editing, it can be less clunky in some ways)

It does exist but it seems to have been disconnected from its other half for some reason.

No, but in this particular case I wouldn't expect the changes to be controversial in any way, and I endeavoured to check as many of the changes as possible (although this was difficult due to the large number of country relations that were in the dataset). None of the changes impact rendering on the main osm.org map, although I would expect them to significantly improve rendering on localized maps like those of Wikimedia Maps.

• Who would I be discussing this change with? Most of these tags were added by a very small number of users who (possibly incorrectly) imported names from Wikipedia or their own data tables, or were added because of the iD language code bug (see below).
• name:xx keys are based on IETF language tags, at least according to the OSM wiki. All of the language codes involved have relevant codes, so I would expect the keys to use those codes.
• I have chosen to capitalize "Hant" and "Hans" because "Latn" is far more commonly used than "latn", and IETF language tags use the uppercase letter.
• Other variants in use in OSM tags: "zh_hant", "zh-hant", "zh-trad" and "zh-traditional"; "zh_hans", "zh-hans" and "zh-simplified". Their meaning is hopefully unambiguous.
• "zh-min-nan" is not an ISO 639 code and is based on the Wikipedia language code for Southern Min. I changed it to "nan". (It is technically difficult to change a Wikipedia edition's language code, as doing so introduces a large number of technical issues that have still not been resolved for the language edition for which this has been done. The Southern Min Wikipedia predates ISO 639-3 by five years, so there was no standard language code for it at the time of its creation. Nevertheless, iD uses the Wikipedia language codes for multilingual names (the issue has been open for almost four years https://github.com/openstreetmap/iD/issues/2457), introducing tagging errors for several languages.)
• "zh-classical" is not an ISO 639 code and is based on the Wikipedia language code for Classical Chinese (I have no idea why these were imported, to be honest). I changed it to "lzh".
• I changed several "zh_py" tags to "zh_pinyin". The latter is much more commonly used and it is impossible to misinterpret the former as meaning something else.
• I made most country codes uppercase. There are very few objects with these tags, and most of them are unnecessary. I opted to uppercase them where they were used in the dataset to match the IETF language tags.
• "Minor" fixes also include key changes for name:pt-br, name:pt-pt, name_zh_classical, and several other keys. I could not use the JOSM validator because of the large number of country relations, although it may be possible by using the reverter on this changeset.

It seems I took a photo of this four years ago (https://www.mapillary.com/map/im/ut3dcq1_TKgK3eazQz62MA). The sign appears to be just before the junction.

Okay. Is the sign that indicates the end of the highway area at that junction point?

In particular, both ways are one-way and thus only accessible from the down ramp.

Are you sure about this? To me it seems more appropriate to use motorway_link on both ways, although I'm not entirely sure if that would be correct.

Multiple name:xx keys are in use to represent the same things (e.g. zh-traditional, zh_hant, zh-hant and zh-Hant for Traditional Chinese), so I have been changing them to the ISO codes, with upper case H to match xx-Latn usage. A while ago I did the same for name:zh-yue to name:yue.

Per changeset/62474451 I've reverted this edit.

JOSM says they changed name=United States to name=United States of America.

Hello, welcome to OpenStreetMap.

Please note that if you're able to add Chinese names, you should add them wherever possible (in Hong Kong the convention is to put the Chinese name and then the English name in the main name field, and then add English and Chinese separately as multilingual names, with keys name:en and name:zh).

Also, you should indicate individual tennis courts if you can draw them separately (see sport=tennis).

Also also, I don't think it's accurate to describe Ocean Terminal Deck as a mall. I haven't been there but https://www.discoverhongkong.com/eng/see-do/highlight-attractions/harbour-view/ocean-terminal-deck.jsp indicates that it might be better labelled a viewpoint (tourism=viewpoint).

I've reverted your changeset because you can't add a website as a physical object. If Questions House has an office or other physical location you can add it with "website=https://www.questionshouse.com".

@BaKaUma the changeset is actually completely erroneous (addition of highway=footway to outline of harbour area), so I'll be reverting this.