All of these cases are somewhat deceptive and deserve more research. In cases 1, 3, and 4 - these areas are slivers or discontinuous areas from actual parks. Case 2 may also be a discontinuous area, but it's not as obvious as the other areas. My suggestion would be to zoom out a little bit and see what's going on around these weird little areas and clean them up so that they represent reality. So the problem isn't really "is this a park or not" exactly, but an issue of scale of the original source data.
On Tue, Oct 1, 2019 at 10:27 AM Frederik Ramm <[email protected]> wrote: > Hi, > > the DWG has been called upon to mediate a conflict between mappers, and > one small part of this conflict is the question of "when is a park a park". > > Some of you know the persons involved and some of you might *be* the > persons involved but I would like to discuss this not on a personal > level and have therefore tried to separate these examples from any > changeset discussions or usernames, and I'm not providing direct links > to OSM either, to avoid clouding anyone's judgement by mixing up > personal and factual issues. > > I have prepared four examples on which I'd like to hear the opinion of a > couple people (if you are one of the mappers in conflict here, please > refrain from participating) but there are more like this. > > ------- > > Case 1: > > http://www.remote.org/frederik/tmp/case1.png > > Two small coastal areas that look a bit like rock outcroppings. I > believe they might originally have come from an nmixter import with a > "zone=PR-PP" which was then interpreted as meaning it's somehow a > "park". It has temporarily been leisure=park AND natural=beach and > park:type=county_park and now it is boundary=protected_area and > leisure=nature_reserve and park:type=county_park and protect_class=7, > without any indication where that protection comes from (and looking at > the aerial imagery it will be difficult to verify anything). > > ------- > > Case 2: > > http://www.remote.org/frederik/tmp/case2.png > > The tree-covered green area in the middle of the image is a > leisure=park, the woodland all together (sharing the eastern border of > the "park" but otherwise much larger) is a natural=wood area. In the > south and west the "park" connects to "residential" areas (that are > partly covered by the natural=wood), in the north the park connects to a > landuse=industrial (also partly covered by wood). > > One mapper says "not a park", the other mapper says that according to > CPAD 2018a and SCCGIS v5 this is a park (none of these are listed as a > source though) and then proceeds to say: > > "It is a park in the sense of American English as of 2019. Whether it is > a park according to OSM may be debatable, as it is an "unimproved" park, > meaning it is under development as to improvements like restrooms and > other amenities. However, it is an "urban green space open to public > recreation" and therefore does meet OSM's definition according to me." > > ------- > > Case 3: > > http://www.remote.org/frederik/tmp/case3.png > > The highlighted area in the middle of the picture straddles a street and > parts of an amenity=parking north and south of the street and seems to > rather arbitrarily cut through the woodland at its northern edge. > > Mapper 1: "This isn't a park. It's just a small fenced off grassy > area.". Mapper 2: "It is a park according to County Park as it meets the > leisure=park definition of "area of open space for recreational use" and > contains amenities (parking)." > > It is currently tagged leisure=park. > > ------- > > Case 4: > > http://www.remote.org/frederik/tmp/case4.png > > Red highlight is a "leisure=park" "zone=PR" (the latter probably left > over from an import). Larger, green area that is mostly overlapping this > "park" but also cutting an edge in the NW is natural=wood. > > Mapper 1: "This park doesn't exist." Mapper 2: "It is undeveloped land > managed by County Parks in a sort of proto park state. How would YOU map > this?" > > ------- > > I find that both mappers here make valid points. Generally, in times > where every teenager maps their back porch as a park in the hope of > attracting Pokemon, I am leaning towards being careful with parks; I > would love to have a rule of thumb that says "if it doesn't have a name > (or if it's not more than xxxx sq ft) then it's not a park, it is just > some trees" or so. Just because an area of a few 100 sq ft is > technically a "park" in some county GIS system, doesn't mean we have to > call it a park in OSM, and the idea that any patch of earth with three > trees on it and two cars parked on it is a "park" because it is "open to > the public" and "has amenities" sounds very far-fetched to me. > > Also, mapping micro-protected areas on a rocky shore seems to be of > limited value to me and puts a big burden on anyone who wants to verify > that. > > But I'd like to hear others chiming in. > > (This particular mapper conflict has other dimensions that just parks > and DWG's further actions towards the mappers involved will not depend > on the outcome of this discussion.) > > Bye > Frederik > > -- > Frederik Ramm ## eMail [email protected] ## N49°00'09" E008°23'33" > > _______________________________________________ > Talk-us mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/talk-us >
_______________________________________________ Talk-us mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-us
