[OSM-talk] New Popular Edition in JOSM

[Andy Robinson]

Andrew Rowbottom wrote:
>Sent: 14 November 2006 9:18 AM
>To: talk at openstreetmap.org
>Subject: [OSM-talk] New Popular Edition in JOSM
>
>Ahoy,
>
>On Sunday 12 November 2006 19:28, Nick Whitelegg wrote:
>> > But now I'm a bit worried. The tracks in my area don't align. (And no,
>> > it's not changes to the street pattern!) If I move the NPE maps to
>align
>> > on a key point, then they quickly diverge in all directions from the
>> > point.
>>
>> I'd be surprised if it's that bad - maybe it's an issue with that
>> particular NPE map. Earlier today I got good alignment in an
>approximately
>> 5km x 5km area on one NPE tile, and 2km x 2km on another.
>
>There was a post on the ordnance maps list on yahoo which quoted a
>reply from the OS, one part which caught my eye was :
>
>> a programme of Positional Accuracy Improvement has been
>> undertaken to address inconsistencies resulting from the recasting of
>> the large scale Victorian County Series mapping onto the National
>> Grid and the Overhaul of these maps during the post-war revision.
>
>I might be wrong, but it looks like this could be interpreted as
>meaning that, in some rural areas at least, the local scale positional
>accuracy of current OS maps is, in effect, based on that achieved in
>the last part of the 19th Century.
>
>If this is true this would probably have been an even more common
>occurrence in early NPE maps. Not being a map historian I can't be
>sure of how big an effect this was, though I  can say with reasonable
>certainty that in the previous series (Popular Edition 1920-30's) even
>after ortho-rectifying on a 2mile grid and reprojecting some of the
>mapped trig pillars can be adrift by up to 200m.
>

This makes an important point about using old maps generally. You cannot
rely on a drawn map with any certainty because there are a number of
contained errors. There are survey baseline errors (what became the trig
pillars and is now done with survey grade GPS), surveyor's errors, and
draughting errors (projection errors are less of a problem on a grid based
map). All of these are likely to be present in the NPE and other old
mapping.

On the whole, aerial imagery (and I include Landsat in this) should provide
a more reliable source for physical location, but as we have seen not even
these are without positional error. What they are good for though is
defining the shape of physical objects, meandering rivers being a very good
example, which are much better defined on aerial imagery even at low
resolution, than is typically found on traditional maps.

I suspect we are going to find in the long run that OSM will be a more
accurate positional map than many maps that have gone before, but that's not
particularly important if the traditional errors are less than a couple of
hundred metres or so. You can still navigate and find your way. What's more
important is that the route is the right shape. This is especially true of
walking maps where I've found over the years that the OS maps have been
woefully inaccurate at defining the true shape of a path, especially when
trying to find your way off some remote fell top in the fog!

Cheers

Andy

Andy Robinson
Andy_J_Robinson at blueyonder.co.uk 


>Andrew Rowbottom
>
>_______________________________________________
>talk mailing list
>talk at openstreetmap.org
>http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk