[Talk-in] Border state | districts
harshad.rj at gmail.com
Wed Dec 16 11:18:12 GMT 2009
Micha and I had some conversation about this off-list and I requested that
it be discussed here.
My understanding may not be complete and Micha obviously has more knowledge.
Here are my thoughts anyway...
The Relation approach seems to be the new way of doing things, but from what
I could gather there isn't a complete consensus on it.
On the one side, it reduces the burden on the renderer to deduce relations
between boundaries and areas; on the other it puts the onus on the map
editor to manage them.
I personally feel that putting the onus on the map editors is not the right
way, cause of the effort involved and because humans are more likely to make
mistakes. In the long run, implementing a better renderer + relation deducer
will be better.
In the short term, if I understand correctly, the only hitch with the
current approach is that borders will overlap.
On Wed, Dec 16, 2009 at 2:49 PM, michael lohr <micha.lohr at web.de> wrote:
> we'll need relations of type=boundary to tag the borders properly.
> there's always only ONE way to define a border. always use the highest
> appropriate admin_level, ie. if a district border is a state border as well,
> tag the way as a state border. no names or anything else here, just
> boundary=administratve and admin_level.
> then intoduce several type=boundary relations, one for each administrative
> unit. for instance: if a district border is a state border as well but not a
> national border, then you'll need 4 relations: 2 for the districts and
> another two for the states. the ONE border way is a member of all 4
> hope i meda myself clear, if not see here:
> or look a himachal, i corrected the borders there yesterday.
> regards, micha
> Talk-in mailing list
> Talk-in at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-in