[Tagging] State parks and state forests: specific tagging question, general mapping philosophy

Kevin Kenny kevin.b.kenny+osm at gmail.com
Wed Jul 27 17:34:04 UTC 2016


On Wed, Jul 27, 2016 at 9:10 AM, Frederik Ramm <frederik at remote.org> wrote:

>  I don't see (a) that everything is ready for it
> nor (b) that it would require any downtime. For example I'm running
> osm-carto with hstore and using views so I don't have to modify the
> style; but it hasn't been shown whether that would be a good approach
> for OSM or whether the carto style should be changed to use hstore
> columns directly.
>

It's also a decision that can be deferred. Enabling hstore while keeping
the existing columns would be Mostly Harmless, particularly if the keys
that have identified columns are not duplicated in the hstore. It does have
the effect of making the hstore keys "second class citizens", but is
a lot better than the current approach of denying access altogether to
keys outside a specific, enumerated set.


> There's quite a few people who have changes in waiting that are only
> possible with the hstore extension. Of course there lies a danger in
> that - without the excuse of "needs hstore", we might suddenly find
> ourselves having to cater to lots of niche requests, aka "if there's a
> tag to differentiate X and Y then I want to see that difference on the
> map!"
>

"Niche requests" are an indicator of project vitality. Your statement
comes across as saying, "the lack of hstore provides us with a
convenient excuse not to be responsive to users and contributors."
I hope that isn't what you meant.

It's probably worth noting that in the related discussion of access=permit
that I'm not proposing ever to make a distinction between the types
for the purpose of rendering the main map. I simply want the information
for maps that I render myself. There's a huge difference between
"make a distinction between objects of type A and type B" and
"make a distinction between objects of type A and nothing."
The first suppresses detail, the second suppresses existence.

> That leaves US users in a bit of a quandary, with only a few viable
> > choices: beef up openstreetmap.us <http://openstreetmap.us> to be the
> > public face of the project on this side of the ocean (a disaster from a
> > marketing perspective, to have two competing faces)
>
> I don't think so. In fact I would like to see more regional diversity in
> "faces", instead of everyone trying to cram their national specialities
> into one central mapping style.
>

I agree wholeheartedly. It's important to note: that sword cuts both ways.
There's a fair amount in the way of Eurocentric (and, more specifically,
UK-centric) specialities baked into the current system. Unless handled
delicately, the whole localization issue comes across as relegating
non-European communities to their respective ghettos. But yes, we
do need maps better adapted to national and local conditions.

> I'm sorry if I'm prickly. I'm frustrated.
>
> I think you're just too impatient.
>
>
You're right. It's only about three years that the discussion of "hstore on
the central server" has been going on sporadically, during all of which
time I've been running an hstore-enabled rendering chain on my personal
tile server. I suppose it isn't reasonable to expect something like that to
happen in less than a decade.

I think that some people fail to comprehend the scale of the problem
over here. Without "tagging for the renderer", virtually nothing that you
see on https://www.openstreetmap.org/#map=8/43.121/-74.539
would be visible: the major highways, city names, and waterbodies
would remain, but everything else would be gone. Not "badly rendered",
simply absent. And that's the case all over the continent -
https://www.openstreetmap.org/#map=7/37.003/-110.814 shows what it's
like in the West. The National Parks, National Forests, BLM lands,
and so on are our administrative disticts in the rural US. Asking
to render an administrative boundary that encloses an area the size
of Slovenia, if not Belgium - as some of these areas do -
doesn't seem to me to be too unreasonable a request.

I also note that it isn't just a US thing. Calling an "area of outstanding
natural beauty" boundary=national_park, or a "regional park" or
"marine protected area" leisure=nature_reserve is just as much
tagging for the renderer as using one of those tags to label a
National Forest, a state park, or any one of the other legal zoo
of protected areas that we have over here, and yet I see such things
all over the map of the UK. It's not a lie, exactly, quite. Those are all
areas set aside to protect some aspect of nature. It's not quite
as precise tagging as boundary=protected_area with an
appropriate protect_class, but it seems to be impossible for even
the Britons to resist tagging for the renderer to at least that extent.
I see that they have the protected_area tagging in place on most
if not all of those areas, all ready to go when and if the renderer
supports it. I do that as well, on the areas that I've edited recently.

I'm not cleaning up the mess, exactly, quite. I'm just replacing
it in my area with a smaller and better defined mess. I'll settle
for that for the moment. It's not perfect.. It's unquestionably still
untidy and mistagged, but it's a lot better than what will result
if someone reverts it.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20160727/7018517d/attachment.html>


More information about the Tagging mailing list