[OSM-talk] Area-type objects and ways along its boundaries

Frederik Ramm frederik at remote.org
Wed Jul 7 12:21:11 BST 2010


Hi,

M?rtin Koppenhoefer wrote:
> so I it seems your country is doing things just as any other country
> and you should think about not glueing borders to features. 

I think glueing borders to features is perfectly all right; I even 
advise people to use roads or river centrelines in boundary relations if 
it makes sense. I think in many cases the boundary definition does not 
exist independently of the feature.

> I come
> back again with some point already mentioned above: additional points
> of roads for attributes like width or maxspeed, etc.: those will all
> "curve" the road slightly (because points are never 100% linear in
> OSM, or keep to be linear after some edits), hence curving the
> boundary as a result.

It all depends on how the mapping has been done. If the boundary has 
been imported from another source that may be one thing but if the 
boundary has been mapped normally then it is very likely that any future 
refining of the road geometry actually benefits the boundary rather than 
making it "less precise".

I know both sides to this argument and they have been discussed 
endlessly. Both sides have merits. It is good to know both techniques 
and use them where they work. Neither technique is superior.

Bye
Frederik




More information about the talk mailing list