# [Tagging] Simplify building:part areas

Javier Sánchez Portero javiersanp at gmail.com
Fri Aug 18 08:36:16 UTC 2017

```Sorry, I should have taken time to give some examples. Please read below (I
rev.

2017-08-18 1:30 GMT+01:00 Clifford Snow <clifford at snowandsnow.us>:

>
>
> On Thu, Aug 17, 2017 at 4:12 PM, Javier Sánchez Portero <
> javiersanp at gmail.com> wrote:
>
>>
>> I am thinking in ways to reduce the complexity that introduces the
>> mapping of parts of buildings. For example:
>>
>
I have reversed the order of the points

* In the wiki [1] says that the outline should be tagged with
>> building:levels and height, but this, if the parts cover the whole outline,
>> is a duplication since these tags will always be in some of the parts.
>> Could I delete the part(s) whose labels match those of the outline?
>>
>
> If you use a multipolygon, then the multipolygon would contain the levels
> and height.
>

I'm refering to 3D modeling of building height and levels, according to
[1]. For example, this building [2] have two heights and should be drawn
two parts inside the building footprint, one with (building:part=yes,
building:levels=1, height=3) [3], and another one with (building:part=yes,
building:levels=2, height=6). As the building footprint [2] could have the
levels and height tags I put them in it avoiding to draw one part. I meant,
the building area is not entirely covered by building:part areas. All the
building in this village was drawn according to this.

I take the rule to put in the building:levels and height tags of the full
building those of the level wich parts sum a greatest area instead of the
maximum values. For a example see the adjacent building to the left [4]. It
have (building:levels=1, height=3) instead of the maximum values
(building:levels=2, height=6) of the building:part [5]. This way I avoid to
draw two parts inside the building. I consider that the maximum
building:levels and height could be calculated by a consumer from the
building and its parts instead. I'm wrong with it? But it's against what
says the wiki [6].

> * If one part is inscribed within a larger one, can I use simple ways
>> overlapped and leave to the render decide how to draw them or should I
>> create a multipolygon for the larger part with the smaller part with inner
>> role? I'm prone to the first.
>>
>
> An example would help. If the building has an inner court yard, then a
> multipolygon would be appropriate, with the inner court yard with an inner
> role.
>

I'm not referring to buildings with holes but to nested building:part
areas. Consider this building [7] with a big one-story part and a smaller
two-story part [8] within it. If I use the full detailled schema I will
need a multipolygon relation for the one-story part, but I avoid this
putting the tags in the footprint (violating the rule of maximum levels and
height in it). I don't have real example at hand, but supposes another
three-story part inscribed inside the two-story part [8]. should I use a
multipolygon for the two-story part to fully separate it area from the
three-story part area? Or could I just draw the inner three-story part,
overlapping both areas?

[1] https://wiki.openstreetmap.org/wiki/Simple_3D_buildings
[2] http://www.openstreetmap.org/way/459549932
[3] http://www.openstreetmap.org/way/459550128
[4] http://www.openstreetmap.org/way/459549958
[5] http://www.openstreetmap.org/way/459550129
[6] https://wiki.openstreetmap.org/wiki/Key:building:part
[7] http://www.openstreetmap.org/way/215569626
[8] http://www.openstreetmap.org/way/459573978
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20170818/8da711d2/attachment.html>
```