On 2018-07-24 23:51, Martin Koppenhoefer wrote:

> We should not remove the details, and nuances in this field, data consumers 
> can deal with it, they will either treat all/most buildings the same (so it 
> doesn't matter to them anyway), or they could be specifically interested in 
> generalized types they can now define as they need, or they are really 
> interested in different dwelling typologies and their spatial distribution, 
> and are happy with what they find in some places in osm. 
> What would IMHO make more sense are lists or better structured trees that 
> show the system / hierarchy of the building values that are in use. The 
> current flat list does not do a very good job in explaining the system nor 
> for finding specific tags.

I am always in favour of initiatives to increase the structuredness of
the data. But we must also not be tempted to force multiple concepts
into a single tag hierarchy. Before we start down that path, let us be
clear what the hierarchy is intended to represent, and what factors are
in-scope (a different brick colour will not lead to a different building
type, but brick-built vs. wood-faced may impact the type). Lets be
explicit about whether it is as-built or as-used, and how to handle
mixed-use buildings. If we look first at as-built, we will need a
parallel tagging taxonomy for the usage aspect; and the other way around
of course.
Tagging mailing list

Reply via email to