On Wed, 15 Oct 2008 19:24:18 +0200 (CEST),
Roger Bivand <Roger.Bivand at nhh.no> wrote:
[...]
No, not at all. The coordinates returned by map() if projection= is used
are only for plotting, are in arbitrary units, and are only documented in
code. map2SpatialPolygons() assumes that the data are in geographical
coordinates, and ought perhaps to guess an ellipse and datum based on the
age of the dataset (world possibly WGS84, US counties probably NAD27), but
allows the user to set the CRS object directly. The projection= argument of
map() makes the function call mapproject() with the given projection on the
data flowing through, I believe.
If you do str() of map(projection=) output, you will see the projection
name used, but note that the coordinates are arbitrary.
Thanks a lot Roger, so here:
se.baffin <- map("worldHires", xlim=c(-70, -60), ylim=c(62, 69),
resolution=0, interior=FALSE)
or in any of the world* databases, can we safely assume "+proj=longlat
+ellps=WGS84" (possibly with "+over" in some cases) in the returned
object? This would be good info to have in ?map.
One problem I also run into with map2SpatialPolygons() with the above
map object, is that the ring is not closed.