On Wed, Sep 29, 2010 at 10:58, Frederik Ramm <frede...@remote.org> wrote: > Peter Körner wrote: >> >> Imported data is dead data - there's no one that feels responsible for it. >> Imports can kill community and give "newcomers" the feeling that there's >> nothing more to be done. Imports *can* help osm but they can also *hurt* >> osm, because osm is about people, not data. > > +1 > > Instead of importing data, data should be mixed in at the rendering stage. > > Importing data can be acceptable if the original source has ceased to update > the data *and* OSMers have a good chance to improve the data, but even then > there is a risk of damaging the community or preventing one from forming. > Imports run the danger of sacrificing the medium- and long-term success for > short-term gains. > > Importing data is, in my eyes, almost never acceptable if the original > source is still maintaining the data (because keeping stuff in sync is > practically impossible and it is better to simply not have that data in OSM > and rely on the original source exclusively), or if imported data cannot be > maintained by mappers (either because there are none, or because the data is > not visible on the ground). > > There may be some exceptions where e.g. importing a basic road grid can > kick-start development, or where having the data in OSM to align other data > with it is extremely usable, but these are really exceptions.
The problem with mixing it at the rendering stage is that it only works for self-contained datasets like contours, but most of our imports get *merged* to existing data, their tags updated etc. _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
