Corey: The Kamploops example was to validate the change detection process
and import. The boundary cutting was arbitrary. I think we can import by
bounding box of 0.5 degree for example and make sure to have overlaps. That
way I would have catch the error you indicate.
Richard: The hydrographic layer (NHN) is better than vmap0. In term of
scale it goes from 1:20K to1:50K. Although there is no geometric
integration between NRN and NHN, the accuracy allows pretty good
cartographic integration.
Yukon is fine for me for beta testing.
Do you suggest to import boundary layer first ?
cheers,
Michel
On Nov 21, 2008 6:31am, Richard Weait <[EMAIL PROTECTED]> wrote:
On Fri, 2008-11-21 at 03:21 +0000, [EMAIL PROTECTED] wrote:
> I think the idea is not to replace existing OSM data. Therefore the
> import should be able to insert only new data. I do not think we
> should exclude areas. We should detect the changes instead. Richard
> proposed to start where osm data is not present. I agree with him.
I believe that the TIGER import started in sparse areas as well.
http://wiki.openstreetmap.org/wiki/TIGER
There may be an advantage to starting with the boundaries first and then
the hydro network (surface water). We have very few boundaries to merge
with, and I expect the hydro data to be much better than the vmap0 and
aerial tracing. Let me know if I'm mistaken.
In an earlier post, Michel suggested looking at the more-complete road
data first, to allow other provinces to catch up. I concur. In fact,
starting with Yukon, with street and place names, and block face
addressing will be a wonderful demonstration. It would be wonderful to
be able to implement the Karlsruhe addressing scheme during the
import.
Best regards,
Richard
_______________________________________________
Talk-ca mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-ca
