[Tagging] Tagging of State Parks in the US

[Kevin Kenny]

On Sat, Jul 27, 2019 at 8:24 PM Joseph Eisenberg
<joseph.eisenberg at gmail.com> wrote:
> We should discuss this at Github to get other ideas and opinions. In particular, please make a well-reasoned argument for why we need to supposed boundaries tagged on closed ways at #3785 - try to be concise and objective.

I'll work on that one. The best argument is already presented - that
it's well-established tagging, recommended on the national_park,
aboriginal_lands, and protected_area Wiki pages, and that there are
about 3,000 closed ways in North America that are tagged with
national_park, aboriginal_lands, or protected_area and not loaded in
the polygon table and hence unrenderable in this scheme. If that
argument alone cannot carry the day, then I despair of my ability to
construct one that will convince the decision makers.

> I think it may be difficult to get protect_class=21 rendered, unless the tag is more precisely defined. While you are using this tag specifically for recreation related protected areas, the current wiki page says that it can be used for
>
> “Community life: religious, sacred areas, associative locations, recreation”
>
> These should not be rendered in the same color as natural protected areas, and the difference between religious area and recreation is very large, so I think these would need different tagging and rendering differently.

Clearly, you are right that nature-protected areas and recreational
areas should not render alike - in fact, an inner highlight in a color
more like 'leisure=park' or 'landuse=recreation_ground' would be more
appropriate for recreation areas.

The argument that religious and sacred areas, and associative areas
(I've never been able to get an explanation of what these are supposed
to be!) should not render the same way has some merit. However, at
present there is exactly one religious-protected area tagged - a
church in Ukraine that would probably be better marked
landuse=religious. (Unless it's some sort of historic site - I'm
clearly not familiar with the locality!)

> 3 options:
>
> 1) make a proposal to redefine the meaning of protected class = 21 to mean recreation only, then we might be able to render it if others use the tag frequently
>
> However, this would not match the IUCN classes that were originally used.

That actually might work.  IUCN defines only classes 1a - 6.  All
others are an OSM invention, and virtually unused.

> 2) make a proposal for a new protect_class - maybe 28 - specifically for recreation areas, such as National Recreation areas and some State parks
>
> This would make a clearly defined tag without changing the existing meaning of protect_class=21

I can count on my fingers the protect_class=21 objects that are in OSM
and do not designate recreation areas. There's very little existing
meaning there. I seem actually to have been the only one to notice
that the class was available and use it extensively. (Shame on me for
not consulting taginfo, but on the other hand, every other tag that
I'd tried on these features got me a scolding from someone!)

> 3) create a new tag, eg
> a) boundary=recreation_area
> b) boundary=state_park
>
> boundary=recreation_area or so something similar could be used for any protected area without requiring the use of multiple tags, and has the benefit of using plain English rather that a random number (I never can remember those), but it would be important to clarify the difference between this and leisure=recreation_ground and leisure=park.

The problem here is that the typical State Park is multiple things:
it's got some recreation_ground and some nature_reserve and some park
and maybe also camp_site and possibly amenities like museums, inns,
concessions ...  It's a mixed land use, united only by its name and
protection. That, at the root, is what has caused so much controversy
- there are too many people who insist that any tag that describes a
land use must describe a single thing, leaving nothing to tag the
State Park boundary and its name. (Tagging the different land uses
within the park fragments it, and the name of the whole has nothing to
attach to.) We also have plenty of things that aren't state parks, run
by local governments and NGO's, that also have a similar structure, so
'state_park' wouldn't be a good choice for the name. But because of
the fact that 'protected for the value to the community of
recreational land' is a common attribute for all of it, the
protected_area tagging seemed like something that most could accept.

I'm trying to use protected_area to head off into a new direction, and
in fact the protect_class=21 is such a green field (pardon the pun)
that it looked promising. It's got the distinct advantage that it's
not a falsehood - these are areas, protected in perpetuity, for
recreational use.

Another problem with drawing a distinction between nature_reserve,
recreation_ground, and park (and throw national_park into the mix for
good measure!) is that I've already seen far too many heated arguments
among mappers on Wiki talk pages, Github threads, mailing lists, and
so on about what the boundaries of those tags are.  Adding more words
to the definitions DOES NOT HELP, because we are dealing mostly with
the fact that trying to explain the concept of 'State Park' in UK
English, particularly to a non-native speaker who lives in a locale
that doesn't have any facilities remotely like that, is like trying to
explain the concept of 'smoke' to a fish. 'Public-access land
protected in perpetuity by a state-level government for recreational
use' seemed to be the broadest possible common attribute that could be
explained.  GIven the heat of the arguments, I don't think that the
difference among those other land use types *can* be clarified.