[Talk-ca] Aboriginal Lands
richard at weait.com
Mon Feb 13 15:03:56 GMT 2012
On Mon, Feb 13, 2012 at 9:48 AM, James A. Treacy <treacy at debian.org> wrote:
> On Mon, Feb 13, 2012 at 08:37:17AM -0600, Tyler Gunn wrote:
>> With wooded areas and lakes I've noticed we tend to just leave them
>> un-merged. I can imagine for boundaries we'd like to have them
>> merged, especially considering they'd be spanning many tiles.
> This begs the question: are there any reasons to merge areas like
> wooded areas or lakes that are broken up?
> Any reason NOT to merge them?
For wooded areas and other natural features, I think this is largely a
cosmetic issue. The argument of "should a wooded area be split in two
where a road passes?" is an old one, and unsettled. Also the original
digitization of wooded areas may seem arbitrary when compared to
more-recent aerial imagery or local knowledge. All of this means that
you would have to be crazy to expect a sensible answer to "how many
wooded areas are within $boundary"
So, for the above natural features, I'm happy to see mappers do
either; leave them separate or merge them. As long as mappers don't
edit-war over it. :-)
For admin areas I think that the situation is clearer. A data user
would expect a sensible answer to "how many cities are within
$boundary?" Having two or more admin boundary polygons, adjacent and
equal in admin level and name would provide an unexpected result if
that bifurcation was the result of an arbitrary NTS boundary.
So, I prefer admin boundaries to be merged if they are split by
arbitrary NTS boundaries.
This is not an impeachment of the CanVec product. It is just a matter
that we have to resolve to use CanVec effectively in OSM. If one
"cares" sufficiently to include a municipal boundary from CanVec in
OSM, one should care enough to include it in the best way that you
know how. :-)
More information about the Talk-ca