[Tagging] Disputed area

moltonel 3x Combo moltonel at gmail.com
Mon Jul 20 23:31:44 UTC 2015


On 20/07/2015, Greg Troxel <gdt at ir.bbn.com> wrote:
>
> Warin <61sundowner at gmail.com> writes:
>
>> On 20/07/2015 1:08 AM, Greg Troxel wrote:
>>> So perhaps a relation that carries the border tag with two ways as
>>> members.  The relation would have the boundary tags, and also a disputed
>>> tag of some sort listing the set of countries involved in the dispute.
>>> Then each member way has a tag of which country (countries really, but
>>> only those adjacent) thinks that way is the border.  We could require
>>> that the ways making up the relation make up a closed area,
>>>
>>> This could get tricky for 3-way or more situations, but it seems
>>> reasonably straightforward for the described case.
>>
>> I'd use the tag
>>
>> source= ?
>>
>> That does simply state the source of the information... why add another
>> tag that does the same thing?
>
> It's not about the source (which is how you found out a fact).  It's
> which party is asserting something, whcih is part of the fact.
>
> Also, the point is to be machine parseable.

How about:
* Map each boundary as that boundary's country sees it, allowing
overlaps. So the France boundary relation is according to France's
views, and vice-versa for Italy.
* Create a relation containing the boudary relations as members, with
roles litteraly set to either "opinion_a" or "opinion_b", and the tags
type=dispute, dispute:opinion:fr=a, dispute:opinion:it=b,
dispute:opinion:united_nations=a, dispute:negociations=peacefull (not
suggesting that the UN either sides with France or is the sole
pan-governmental organisation whom OSM should tag the opinion of).
* Setup a QA looking for overlaping boundaries without an acompanying
dispute relation
* Work on renderings to take this into account.

There's no need to add a disputed=yes tag to the boundary relations
themselves, the fact that they're members of a dispute relation is
enough.

While that dispute relation sounds nice to me, it may be a bit naive.
It really needs a review from people familiar with various disputes
and with potential data consumers such as renderers and geocoders. For
example, what if a country has multiple disputes with neighbouring
countries ? Does puting the boundary ways instead of the boundary
relations as members of the dispute relation work as well ?



More information about the Tagging mailing list