[Tagging] Why using place=city for legal status is a bad idea

Liz edodd at billiau.net
Sun Dec 27 23:03:38 GMT 2009


On Mon, 28 Dec 2009, wynndale at lavabit.com wrote:
> A sentence was recently added to the wiki
> [http://wiki.openstreetmap.org/wiki/Key:place] stating that the tag
> place=city should be used for legally defined cities in countries where
> such a status exists. This is superficially attractive but it gives rise
> to problems. The key issue is that place names, like roads, naturally form
> a hierarchy, and that the tag place=city is similar to the tag
> highway=trunk.
>
> For instance, the category in the United Kingdom includes Wells, which has
> a population of 10,000 and which the Collins New World Atlas leaves out of
> its 1:3,000,000 map of the British Isles. This is similar to the road
> category of Bundesstrasse (nationally maintained road) in Germany, which
> includes some quite small roads while omitting ones that are much more
> important. If (say) highway=trunk was used for such roads a map generated
> only from OSM data without an outside cheat sheet gives a warped idea of
> roads in Germany; OSM solves this problem by not using a distinct level
> for these roads.
>
> Using tagged population figures as a substitute to guide renderers has
> been suggested but it brings its own problems because people will expect
> them to be more broadly useful and figures may either be out of date or it
> may be unclear what area is being counted.
>

Cartographers use place names to fill in the space. When you to remote areas 
very very small places get a mention because there is space to fill. 
Alternately a compass rose with ornamentation can be used to cover up 
emptiness. 
If you try to enforce a system in which city is defined the same in each 
country you will have the cartographers problem - how to fill the empty space, 
and how not to overfill the space in a place like Japan or Indonesia.




More information about the Tagging mailing list