An update: after some cursory examination of just the sort I am
proposing (USFS data imported from shapefile WITH a "potentially
noisy" transformation with superimposed USGS data), I find very high
data correlation. The largest errors I am able to see (for the small
subset of data I have visually examined like this) are 2 to 3 meters
at most -- these following range/meridian/township lines. Boundaries
which follow creeks or parallel existing, uploaded roads are
essentially spot-on, or maybe centimeter-level errors at worst.
So, "off by a kilometer" may be true as a POTENTIAL error in this
transformation, but from what I see, any error is much less than 1%
of this. I am comfortable with this quite low level of noise in
these data potentially imported into OSM using this transformation.
I will, however, continue to both solicit opinion here and examine
more data, especially at edge latitudes where the transformation may
be more problematic.
SteveA
California
_______________________________________________
Talk-us mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-us
