Mateusz Konieczny <[email protected]> wrote: in reply to the 1 Oct 2019, 16:26 post by Frederik Ramm <[email protected]>:
> Case 1: > http://www.remote.org/frederik/tmp/case1.png > Two small coastal areas that look a bit like rock outcroppings. > It is hard to imagine to me situation where > it would be leisure=park. That is because OSM often uses a definition of "leisure=park" as it is better-known in Europe as (approximately) "smaller urban manicured green space" (and which I tried mightily six months ago to remedy in our park wiki and its Talk page, but was ineffectively confused / muddied by the other party in this "dispute"). We (he and I, the two whom Frederik says are "in dispute") actually DID come to a relatively benign agreement here, in this specific case, where these two polygons are now tagged leisure=nature_reserve (NOT leisure=park) as well as boundary=protected_area (as they are, according to CPAD and/or SCCGIS, which have been documented in our Contributors wiki at https://wiki.osm.org/wiki/Contributors#California_Protected_Areas_Database for many months and https://wiki.osm.org/wiki/Santa_Cruz_County,_California#Landuse wiki for many years) + protect_class=7. The value of this key was actually one of the few times where the other OSM volunteer who disputed this agreed with me that the correct value could be one of two or three different values, and this is the one we settled on. To be clear, the "dispute" that Frederik appears to be arbitrating is over me closing an OSM Note here where I noted this, saying "Resolvable, resolved" and then closed the Note. The other party ignored the Note he left open for months (five months, without doing anything to resolve it, in another disingenuous gesture) to apparently "stir up muck" (vex and annoy, really). Recall, these are from a ten-year-old nmixter import which I spent literally several years and tens of thousands of thoughtful "compares" (the SCCGIS data that were entered, vs. my good knowledge of how OSM should "best tag") and this is one vestige where another volunteer (the other in this dispute) found fault with my corrections, then HE entered a Note (many, actually, this is one), which resulted in the "compromise entries" we find now. I remain in a listening mode as to how other mappers would tag what came from a state agency calling these a "park" and that went through MANY iterations of "not really a park," which I readily and certainly have agreed to (factually, it is not tagged leisure=park today). The Note (https://www.osm.org/note/1759733) was closed, correctly in my opinion, as it is truly "resolvable." By the way (Frederik), I don't know how "if you are one of the mappers in conflict here, please refrain from participating" works in Germany or OSM in general, but I am not used to nor do I appreciate being told to not speak up for myself when my edits are called into question. It feels very much like censorship, having tape placed across my mouth, or having my hands tied behind my back. We have freedom of speech in my country, we have freedom of speech in OSM (so I believe), even during disputes under arbitration by the DWG and especially when some of the facts presented are slightly in error. Yesterday, I meant to send the correction (not really participation, but correcting a mistake in the data presented to talk-us) directly to Frederik, but I mistakenly sent to the list (something I virtually never do by mistake, but I did make that "Reply all" mistake yesterday). Nonetheless, I fail to see the value in Frederick's / the DWG's "ask" that I refrain from participating, especially when some of the facts presented are not quite correct (I place no blame or value judgement about that, it is entirely possible that Frederik's edits and mine simply crossed over eight time zones — a perfectly innocent explanation). I, too, value as many other participants and perspectives as we might view here in talk-us on these topics. They are difficult, they cause friction and I wish to see light, not heat, though resolution (on many fronts, by many volunteers) has proven exceedingly hard to come by. > "zone=PR-PP" which was then interpreted as meaning it's somehow a > "park". > Is this a typical quality of this import? Mateusz, again: "this import" was from a notoriously "import happy" mapper from ten years ago who I know personally and has been widely admonished many times over during the last decade for his poor edits. MY participation was to improve the data into what BestOfOSM.org eventually called "nearly perfect landuse." I have striven to do this over many years, as best I can, logically mapping the imported (zoning) data to OSM's landuse tags, with full explanation of my reasoning all along the way in wiki, personal messages and patient answers with all and sundry with whom I and others interacted as we edited these (multi)polygons here. When others dispute(d) my findings or tagging, I listen(ed) and usually / often / nearly always concur, making corrections every single time. If not I explain my position and document it (in wiki, source tags, changeset comments, Notes or elsewhere), in good OSM fashion. There HAVE been misunderstandings between these landUSE data and landCOVER (even by members of the DWG) — this is a frequent blurry line of confusion in OSM. Nonetheless, I contend that the data in OSM in the county in question are largely if not very nearly completely correct (I certainly say these data are "95%+" correct, perhaps closer to 98% or 99% depending on how you might score). Newer tagging schemes which BETTER tag or UPDATE the tagging are as welcome here, associated with these data, as they are anywhere else in OSM, that is to say: if you have better data or tagging to enter, please, by all means enter them. I listen, I "get out of the way" when people do this, as I am not "wed" to the imported data, though I don't wish to seem them simply deleted without being replaced or updated with superior data. I think that is reasonable, other reasonable people have told me they find this reasonable, too. > Case 2: > http://www.remote.org/frederik/tmp/case2.png > (...) > One mapper says "not a park", the other mapper says that according to > CPAD 2018a and SCCGIS v5 this is a park > Aerial image is useless here, it > is a tree covered area. That is because aerial imagery displays (largely speaking), landCOVER as opposed to landUSE, which is what these polygons describe. You can't (always) see landUSE in imagery. As I said to Frederik off-list recently: "The natural area abuts a mapped residential boundary, effectively becoming 'the backyards of people's residences.' This, too, is a frequent occurrence (parks/natural areas abutting people's backyards), so the existing polygons merely define boundaries between public and private: 'you may park here, you may recreate here, but beyond THIS line (polygon boundary), now you are in somebody's backyard on their private property.' I consider this an important line to map, as I respect private property and don't want to trespass when I recreate." So, in short, using aerial imagery to determine whether these polygons are correct (as landUSE) is a NOT-especially-helpful strategy. > It may be in addition leisure=park, > it may be a dump of nuclear waste, > it may be a military polygon. > > Is there a chance of on ground photo? Again, there are many sources of on-ground photos, but few, very likely none, will describe landUSE as these polygons are tagged. > I am unfamiliar with CPAD 2018a and SCCGIS v5. > Is there a good reason to expect that their classification > matches OSM classification of objects? There was a cavalier and sloppy attitude by the original importer in 2009 about "classifications matching," but over months of my better understanding in 2010 and the years between 2011 and 2014, I believe we largely "got it correct." Newer versions of the data were updated in 2018 and 2019, and IMPORTANTLY, many OTHER data have been entered which SUPPLEMENT and/or UPDATE these data, which again, I welcome and so does our map. This is very much how OSM works: think of how the TIGER data in this country are slowly-but-surely being replaced by superior data: it's the same in this county. > "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. > I would not expect restrooms to > be indicator of leisure=park > However, it is an "urban green space open to public > recreation" > I am one of people that attempted to > improve OSM Wiki documentation > of leisure=park Actually, Mateusz and I participated in some good work together on minor modifications (what I believe both of us thought were clarifications to) OSM's leisure=park wiki, but which it later became clear to many seemed to only further blur the lines of definition. That is unfortunate and we still have much work to do to clear up semantics and dissolve remaining confusion. This will be difficult, there is no doubt about that. Please let us be polite and accommodate that there are many different interpretations of what "park" means (the OSM version, the "US English" version...even others). OSM really must strive to be inclusive "in a very large tent" here, even if this means a wholesale re-working of how we tag parks and similar lands in OSM. (Two other well-respected mappers are participating with me in United States/Public Lands, one of them a serious and deep mapper in New York, the other a fellow California mapper, who happens to be an OSM wiki moderator). These are VERY difficult topics, let's please agree to that as a positive starting point and work forward from there. More-or-less autocratically deciding "these data are crap" without understanding this basic tenet and histories of how tagging has entered and evolved in the last 10 to 15 years is not helpful to the cause. > 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. It is absolutely not arbitrary, and it doesn't display in imagery as what it is: the delineation between public (park-like, actually now better-tagged leisure=nature_reserve) and private lands. To the north are "residential backyards," where recreating on the nature_reserve is not welcome, as it is private property. Again, it isn't simply me who finds these boundaries important and valuable to include in OSM, others who respect private property don't want to trespass while we recreate and we appreciate having these on a map, in the palm of our hands on a GPS, etc. > 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. > Is there a chance of on ground photo? No, again: it is now "better-tagged" leisure=nature_reserve to exactly avoid this sort of dispute. It is a natural area which might be described as a "proto-park," a parcel of land publicly owned, yet unimproved (save for a parking lot and some natural area) with things like "trimmed greenery" or "restrooms, playgrounds and other amenities frequently found in parks." I believe most would agree this isn't a park, but it is an area where people congregate, recreate, is open to the public and largely remains in a natural state. Hence, leisure=nature_reserve is either "fairly close" or quite accurate tagging here. The corollary is that the Note (one of the disputes) could be, and was, correctly Resolved. > I would give low weight to whatever it is officially considered as a county > park > 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?" > Park is not there so I would not map. Why is that? Why does this deserve "low weight" to not only how "we" (the People) designate through our Parks Department (whose job it is to manage land for public recreation) and ACTUALLY USE THE AREA compared to what you think you can see in a satellite photo? I've been here, I park here, I recreate here. Do you? I'll ask you (and all and sundry, as Frederik is) as well, how would YOU tag this area? > Though mapping it as a garden may also work. > 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, > +1 Ugghhh: AGAIN, we are stumbling forward from a messy import, a much-better logical mapping to landuse tagging over many years, the development of the didn't-exist-when-the-data-were-imported schemes like protected_area and likely most important: the "clarity that we are confused" about the leisure=park tag, especially when it is applied in the USA, by people who speak and use US English (I define "park" in this dialect several times in our wiki, and it DOES diverge from what OSM says is leisure=park, which is only part-speculation as to why we have these problems in the first place, not the end-all and be-all explanation for it). If you are going to "+1" stuff, please understand the history and context of how they happened. Otherwise, you may very well be calling something "wrong," when actually it is "not fully and properly evolved from how it originally entered our map, was improved over the years, and newer tagging schemes have superseded its relevance or wider understanding as correct." That IS what happened here. We should not start from the place that people wish to mis-tag or confuse, I don't, I don't think the other party did and I don't believe that there are "bad desires" at the root of this. It appears to be blurry, poorly-defined semantics and getting-lengthy histories of complex tagging, and we continue to unravel those. Yes, it is difficult. > 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. > +1 This is simple hyperbole and therefore not likely worthy of serious consideration. Let us address specific cases rather than "throw the baby out with the bathwater." (An American English idiom roughly meaning to flippantly or carelessly discard something wholesale because it has a slight wrong with it). > 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. > +1 Why? We "micro" map things like schools, telephone/fiber/cable boxes and many other "micro" things, is there a clear reason why this sort of "micro" is "of limited value?" I see no good argument for that here. Why is it a "burden" to verify when the data are pointed to by our Contributors page (or the County page in which the data reside) and as the data are freely available on the 'net, can be downloaded and compared to what is in OSM almost as easy as it is for me to type the words it takes to describe that process? While I welcome other volunteers in our project verifying data, I question when people call mapping these things in the first place "of limited value" and that their mere existence in OSM places a "burden" upon other mappers. And please, as you contemplate that, please understand the (again) frequently confused topics of how many places are predominantly tagged with a landCOVER flavor, while others are largely tagged from a perspective of landUSE. Both are correct in OSM, and you (and others) may simply not be used to seeing mapping in "the other" method of tagging such areas. Actually, I believe one of the better results that may (decades from now, to be candid) emerge from this is how we tag BOTH landUSE and landCOVER either one, the other or BOTH, successfully. That's ambitious, but we might very well get there. Let's remain civil and true to our first name — Open — while we do. SteveA California _______________________________________________ Talk-us mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-us
