I propose to use the ‘height’ tag only for buildings or parts of buildings that actually have this height and not for the superordinate outline. Furthermore, I prefer to specify a building complex from top to bottom. This means that the highest part of the building (unless the others are specified in more detail) is given the name and links etc. and not the outline.
It does not make sense for 2D to summarize the whole building complex with all tags. The outline serves only to delimit the building geographically, not to specify it.
E.g. Church tower:
Only outline, no other parts of the building (outline contains all tags including the height and links …)
Outline includes other parts of the building
a. Outline get only relevant tags, i.e. primarily the geographic coordinates
b. Ship (height, roof, …)
c. Tower (all tags, height, names, links, …)
For future 3D applications it is important that the highest point is named and not the outline, I think that is also an advantage for 2D, because you can then see where the highest part of the building is from the top view. The information should be visible in 3D on the top of the church (tower) and not somewhere in the air in the geographical center of the outline.
If you provide the outline with all the tags, you have to assign the building parts to the relations and outlines again very time-consuming during the data extraction.
You’re considering a single use case here: A 3D rendering where one wants to have buidings’ names hover over the highest point of the building. That might become a little bit easier.
But there are many other use cases where moving the building’s tags to the tallest building part isn’t helpful at all, for example:
Entering the name in a search field and wanting the entire building that’s the result of the search to be highlighted (e.g. glowing).
Clicking anywhere on the building to bring up a popup with information about the building (name, architect, Wikipedia link, …).
And of course there’s the use case of rendering buildings in 2D, which is still by far the most common. That should preferably be possible without considering building parts. (Although I do agree that even a 2D rendering could sometimes benefit from considering building part information – but that should be optional.)
So I think the current approach suggested by Simple 3D Buildings – that the name, links, etc. should be added to the building outline – is ultimately far more practical. Adding information about the building as a whole to the building=* object (the outline) also seems more intuitive, imo.
I think 3D applications getting more and more important and we should consider them more. Both approaches (upd/down, down/up) have their advantages and disadvantages. Nevertheless I think we should use the ‘height’ only for buildings or building:parts which has actualy the height and not for describtion purpose in the footprint/outline. Using the ‘height’ in the outline makes reconstructing the 3D shape nearly impossible, because the footprint has mainly a different height as the building:parts and to avoid redundant data. Beside do know a tool or process for c++/win10 to extract the building shape based on the planet.pbf?