On 19-Jun-17 08:30 PM, Martin Koppenhoefer wrote:
2017-06-18 17:44 GMT+02:00 Joachim <nore...@freedom-x.de <mailto:nore...@freedom-x.de>>:Currently we categorize settlements by number of inhabitants, modified by importance. I would rather not open another dimension. The type of of a settlement can be deduced by processing the landuse inside. If there is use in a binary (yes/no) property it should be added to the place=* feature: e.g. dispersed_settlement=yes. But this looks more like job for Wikidata.we base our settlement classification not only on population, the population numbers serve as an indicator what size we expect from certain classes in certain areas, but they are not hard limits. Other factors we consider: part of a settlement or whole settlement, functional/cultural parameters (a place which is officially a "town" in Italy will be a town in OSM, a place which isn't a "town" will not become more than a "village") and maybe more.
The functional aspect counts to a fair degree in Australia. As the over all population density is low a small remote settlement will usually have more facilities that a European would expect from the settlements population size. Settlements can serve a vast area with high demand for these facilities. If a settlement failed to provide these facilities travelling times to obtain these services would be large (days there and back).
_______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
