On Sat, Dec 13, 2008 at 1:45 PM, Sam Vekemans <[email protected]> wrote:
> D -are we in unanamous agreement that the way to do the import is to; > Create a program which will detect existing OSM data and add the UUID# > reference onto the point/way/polygon/relation so that it will not > interfer with existing data, placing only new points/ways/polygons and > so future updates will be able to use this software to only add new > data.? The chances of finding a node in the OSM database that has identical coordinates to the Geobase data is going to be pretty slim. Finding a way/polygon/relation with identical information is going to be several orders of magnitude more difficult. The issue that I see is trying to match existing data in the OSM database to the imported data. Using algorithms that can match nodes/ways that are within a specific tolerance value, and thereby merge the two sets of data leads to another issue. Do we use the OSM coordinates, or the Geobase coordinates? It could be that the OSM data is of a higher quality (more accurate) than that provided by Geobase. If we use the Geobase data, then on subsequent Geobase updates, it would be easy to match the ID and coordinates. However what do we do when the ID matches but the coordinates don't? ie. An OSM user has moved a node, or possibly the Geobase node location has been updated? >From my point of view, it looks like whatever we do, there is always going to be a lot of user intervention when a mass update/import happens. Two separate databases being maintained by unique user groups is always going to be in conflict. James VE6SRV _______________________________________________ Talk-ca mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-ca
