The reasons *not* to use the Zillow dataset are clear: nobody but zillow can edit it, and it is based on low quality TIGER data.
The flickr dataset is similarly suspect, if this is any indication: http://boundaries.tomtaylor.co.uk/#23512042 It shows San Francisco's "*SoMA*" (South of Market) extending north of Market. And Mission Bay extending past Mission Bay. And.. and... a thousand other problems. --------------------------------------- OSM is just not a good place for opinion: and neighborhood boundaries in many many places are a matter of *opinion*. I'd say think like a map user. Clearly you want to be able to enter "*SoMA*" into a search engine an get at least some idea where it is. Using nodes for neighborhood names hits a large fraction of the use case, with a small fraction of the problems of polygons. I'd say that OSM should have neighborhood names as *nodes*. Then conflate or link to other databases that have richer data. The OSM node could even link to a wiki page where the neighborhood can be described in all its richness and complexity. *OpenNeighborhoodNotes.org?osm_id=23424234* anyone?
_______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us