[Talk-us] Parks in the USA, leisure=park, park:type

[OSM Volunteer stevea]

On Apr 24, 2019, at 2:05 PM, Kevin Kenny <kevin.b.kenny at gmail.com> wrote:
(a LOT about parks!  thanks, Kevin!)

> TL;DR
I tried to be brief, sorry if I wasn't.

> - Tag the land use, not the land ownership. A city, town,
> county, or state park may be virtually indistinguishable urban green
> spots, recreation grounds, nature reserves, whatever. The level of
> government that manages them may be of interest and worth tagging, but
> ought not to be the primary determinant of 'park type'.

I tag a whole heck of a lot of land USE, yet exactly HOW do I tag a typical "county park?"  (Mmm, there is nothing "typical" about these).  This is what we 'mericans largely call "park" yet doesn't hew to OSM's newly freshened-up leisure=park, which now more strictly means "smaller manicured urban public greenery, shady, tidy, semi-natural places to walk within the city, likely a restroom, maybe a playground..." with the emphasis on "smaller" and "urban."  County parks are often more-rural and can be quite large.  Accordingly, the newly-narrower leisure=park tag seems no longer an even somewhat-correct tag on these.  So what IS the "land use" here?  Especially when it clearly ISN'T leisure=park?

I do not mean to put as much emphasis on "level of government which administers the park" as people take here:  it's almost a non-issue and can be fully captured by operator=* and/or owner=* tags:  if they better clarify, use these.  The park_level tag is an old idea of mine we might resurrect to aid in better rendering park boundaries if we so choose, that's all it would be good for, same as admin_level acts today.  (There are places, especially in far northern California, where visually parsing the cacophony of different park jurisdiction boundaries would greatly benefit by semiotic aids to do so).

> I think that the Wiki definition leaves a lot to be desired, and I'm
> groping in a fog, much as you are, so please don't take anything that
> I say here as a confrontational pronouncement.

I'm glad to hear you grope, too, as I know you've had a lot of interaction with these taggings and what might be done about them.  As I've said, it's a chewy problem.

> My read on "urban/municipal" is that it describes setting and land
> use, rather than the operator. To me a "park" in a
> urban/suburban/front-country setting connotes a certain type of
> facilities. It will likely have adequate parking, or else access to
> public transportation. It will likely have public toilets.

Right, this is what I meant by "admin_level=8, LARGELY" as leaving that wiggle room is truly required:  it isn't ALWAYS the city parks department that will operate every single leisure=park in a given city.  Still, look at how vague is talking about "setting."  That's difficult to agree upon right out of the gate.  (I'm not complaining, merely re-stating the difficulty of articulating the problem, even as we do our best to tease out what we mean).

> ...these features make for what is essentially a human landscape...definitely human-sculpted.

This is a potentially excellent addition to the leisure=park wiki, as you do capture an important semantic with this.  Thank you.

> A 'national park' ... (is contradistinguished) to...the rest of the zoo of
> NPS-managed facilities)

But you actually seem to glom them together because of their many similarities.  I agree these seem much more similar than they do different.  Still, we are left with "national parks" (and things which are so much like them that the tag might fit well, more-or-less), leisure=park (which we agree "we know them when we see them," yet are hand-wavy vague beyond what we now say in its wiki) and this great big slew of "other things called parks" which largely happen to be things like county parks, county beaches and similar ilk, which do NOT fit (neatly or otherwise) into those two categories.  Hence, the conundrum continues.  Especially as I ask again, what IS the "land use" on these?

> It's common for large 'parks' (suitable) to introduce beginners...

This is (almost?) yet another category of (loosely stated) "park," perhaps "a kind of human recreation area" which perhaps we have yet to well categorize and tag thusly.


> 'Nature reserve' covers a lot of things...particularly in North America

It does seem N.A. does things differently than others in OSM and the greater world, but it may be that I simply haven't done enough homework or traveling to fully and more correctly state that.  This (parochialism, regionalism) may be a primary source of our difficulty.  (I have been to three continents, but of course I haven't been to nor do I know everywhere — I more and more rely upon OSM for that!)

> ...forests and (effective) game reserves.

Thank you, this offers crucial knowledge which definitely should be expressed in precision OSM tagging.  I know you do your best to achieve that where you map.  We should all strive to do so well at tagging, which is what many see as a topical holy grail.

> I think it's a fine idea to have park_level (although given that we're
> also talking about nature reserves and forests, that's an unfortunate
> name) so that we don't continually confuse the type of resource with
> the level of government that manages it.
> (and some cool history).

I'm all for another name if you'd like to coin/suggest one.  However, you say "the resource" and that's a good (at least temporary) placeholder word to plug a hole I've mentioned:  what to call these lands which seem not to fit into the existing ones.  We know national_park, we know leisure=park (and there are also leisure=nature_reserve and boundary=protected_area, too).  Call these the "Existing 4."  Yet, (pulling the bubble gum from behind my ear and blowing), we might now also have "public resources" (as land, sometimes water — "area," really) of whatever level of government, that's a minor concern, if one at all.  Perhaps I'm looking for a new key to tag with, call it public_resource, into which all of these entities can live together as one big semantic bucket, the elusive "5th key" into which all of today's question marks magically belong.  It would be great if we could agree on one, AND have it render in Carto relatively quickly, that would solve a lot; that might be a central nut to crack here.  Or, maybe we don't need a 5th key.

Are county parks leisure=park or boundary=national_park?  No.  Are they leisure=nature_reserve?  Sometimes, if so, use that tag.  Are they "always otherwise" boundary=protected_area?  Well, maybe, I don't discount that possibility.  Maybe there is a need for a public_resource (or whatever) new 5th key, maybe not and protected_area covers them all, I truly don't know.

> Then we can actually work on trying to sort out park types...

...with what feels well-thought out, yet I can't help but think it is only thought out as far as you could go today (which is great), yet it will run out of gas sometime in the near future, as it wasn't FULLY thought out.  Defining today's syntax to describe today's semantics does have a tendency (especially in OSM) to not quite fully define today's semantics, and it certainly doesn't describe tomorrow's semantics.  Again, tough stuff to do well/right.  We do our best, tagging evolves, that's the reality of this.

> The 'type' of park ought to be "what type of experience ought the
> visitor to expect" and not "what government manages it."

I say "yes" to "not what government manages it," as I repeat, that's largely a sideshow (though "county parks" do seem to be a frequent bucket containing things which fall through the cracks of the Existing 4).  Yet I almost want to say "no" to restricting us to the Existing 4, as those four might not be enough.

> ...Robert Moses State Park https://www.osm.org/relation/6442393 is a swimming beach. I
> see no reason to map them as anything but what they are, except to
> inform the user that they're state-owned and run by OPRHP (the Office
> of Parks, Recreation and Historic Preservation).

Well, there's some fairly complex tagging here, some may take issue with the use of both leisure=park and yet protect_class=2 (and protection_title — documented — with protection_object — undocumented) without the key of boundary=protected_area.  I've never seen that methodology, but I don't disagree with it (meaning I find it correct, if a bit chin-stroking curious).  They are nicely tagged with operator=*, website=* and phone=* tags, too, so, good.  Yet this fits rather nicely into what many would call an urban, leisure=park, even though it is a state park and might receive a tag of boundary=national_park for exactly that reason, yet isn't.  It seems this exemplifies your desire to see "what it is" tagging, which I certainly respect and I believe is your point.

I'll continue to mull this over, as I thank you for your prodigious contributions to the discussion and my resulting (somewhat better) understanding.  The conundrum continues, the fog is still there, though maybe it has lifted a bit with "Tag the land use," which I already know to do.  And I am better untangling the "government level" as perhaps largely irrelevant (well, it should be tagged when known) from "what is" tagging.  That's a helpful step in the right direction, as it appears to be a vestige of "where the data came from and how they got into OSM."  That shouldn't matter, really.

Whew,
SteveA