On 04/01/2010 03:06 PM, Tom Hughes wrote: > On 01/04/10 14:42, Kai Krueger wrote: > >> Perhaps even easier and a bigger win, would be to import the postcode >> data. It is "only" points anyway, so many of the aspects making data >> imports hard, such as connectivity and duplication, don't apply as much >> here. Furthermore, given postcodes are unique identifiers, it would be >> very easy to spot which (full) postcodes are already in the database and >> only import those that aren't yet mentioned. > > Are random points that just mark postcodes appropriate? I know people > have in the past added so called "postcode centroid" points but I have > tended to remove those when I come across them.
Did you just delete data, or replace it with more accurate postcodes? If it is good enough for Ordinance Survey, it can't be all that bad an idea and once we get better data through surveying, we can replace the centroids with per building postcodes. The usual way of iteratively improving things when possible. Where do you see the problems? Kai > > Tom > _______________________________________________ Talk-GB mailing list Talk-GB@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-gb
