[Tagging] Urban perimeter

Martin Koppenhoefer dieterdreist at gmail.com
Tue May 27 14:09:53 UTC 2014


2014-05-27 15:01 GMT+02:00 Nelson A. de Oliveira <naoliv at gmail.com>:

> On Tue, May 27, 2014 at 7:18 AM, Martin Koppenhoefer
> <dieterdreist at gmail.com> wrote:
> > For settlements (in urbanism / settlement geographic terms) my
> suggestion is
> > to use place=* on an area.
>
> There is one problem when using place: people start to duplicate data.
> What is already present/available in the place node is duplicated in
> the place area. Sometimes it also causes data mismatch (population
> with one value in the node and a different one in the area, different
> classifications, names, etc).
>


It is not really duplicating data, as the place on the node does not convey
any information about the spatial extension of the place.
This is essentially a relic from the early days of osm, interpreting a node
is easier than a relation (and relations weren't even there when we already
had mapped a lot of places). The nodes convey a different info though, that
of a "central spot", useful for rendering and for generic routing (without
a specific address) or for computing distance tables.

The solution on the long run could be to have place relations to combine
the area with the place node. Duplicating data like population should be
avoided, I agree.

These are all well known issues, but it takes some time to sort this out.

cheers,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20140527/e69c843c/attachment.html>


More information about the Tagging mailing list