Alex Barth <[email protected]> writes: > Hey Serge - > > On Tue, Sep 17, 2013 at 2:51 PM, Serge Wroclawski <[email protected]> wrote: > >> Alex, is there a reason you think it's better if people do it? > > > Going auto will take significant work (essentially building the script that > determines that there's no building in a given area) and I'm not keen on > committing this without really _knowing_ we need it.
The Massachusetts building import did this - I think exactly what Serge's note was asking. The scripts are available, and I think Jason Remillard has sent pointers to them multiple times. We then reviewed town (as people had time) to see if the two osm files (massgis buildings not overlapping anything in the db, and massgis buildings that overlapped) to see if we thought they were right. And then Jason debugged the script. Finally, towns that had been manually spot-checked were uploaded, and then the rest when we were just not finding any real problems in examined towns (just a few stray tents marked as buildings). This was a good way to do it, and if we had it to do over again I'd want to do it the same way. Jason is to be commended for slogging through this and responding to the feedback about the process. Also, for the record, the Massachusetts building import resulted in a dozen or so mapers (a fairly large % of the total by editing weight) being in much better touch by private email. Many of those people don't participate in national- or global-scale lists.
pgp9ZsQ6qwxHK.pgp
Description: PGP signature
_______________________________________________ Imports-us mailing list [email protected] https://lists.openstreetmap.org/listinfo/imports-us
