[Talk-us] Usage of highway=track in the United States

Kevin Kenny kevin.b.kenny at gmail.com
Tue Feb 23 15:09:56 UTC 2021


Paradoxically, the fact that we are still having this discussion shows how
unimportant it is.  If it really made a difference, people would have
already adopted a practice of tagging for the renderer or tagging for the
router. With tracks, service roads, and many residential ways being the
'leaf nodes' in the road network, it really has no effect beyond trivial
differences in rendering. Routers tend to get it right, since most of their
algorithms disfavor service roads and penalize tracks severely.
Essentially, you'll get routed over either only if they go to your
destination.

(Take it from a man who has had a GPS scold him, "you have been driving off
the road for the last one point five miles. Please proceed to the indicated
route.")

Can anyone present me an argument, other than a purely philosophical one,
for why the distinction actually matters?

On Tue, Feb 23, 2021 at 2:09 AM Joseph Eisenberg <joseph.eisenberg at gmail.com>
wrote:

> While I appreciate that you are attempting to document how this tag is
> actually used, there seems to be some too strong assertions in the new
> section:
>
> 1) highway=track is used for paved forestry roads in parts of the USA with
> heavy rainfall, e.g. Oregon:
> https://overpass-turbo.eu/s/141R  - for example
> https://www.openstreetmap.org/way/5520248 and several nearby private
> forestry roads in the Bull Run watershed. There are also some logging roads
> in privately owned forestlands which are paved. There are 265
> highway=track + tracktype=grade1 + surface=paved ways in California:
> https://overpass-turbo.eu/s/141S
>

Don't assume that they are mapped according to common local practice.  As
you point out, there are more than that number of MISmapped tracks in New
York City alone. Saying 'these 261 outliers prove that it's proper' while
arguing 'those 400 outliers are all errors'  is wrapping your assumptions
into the argument. There surely have been US mappers who've followed
European practice, particularly after being scolded on this mailing list.

2) Re: usage "For any unmaintained, minimally maintained, or overgrown
> road, regardless of usage or purpose"
> Can we actually cite examples of ways roads that meet that standard but
> which are not primarily tracks from a usage standpoint?
>

I know of a few, like the abandoned County Route 3 in Lexington, NY - where
I'm told it must be either 'service' or 'residential' because there's a
cabin or two on it. Find it at https://www.openstreetmap.org/way/291345321
It's a road without public maintenance for automobiles but that the public
has a right of access. It's mostly used for snowmobiles and ATV's although
once in a while they come in and cut the brush on it so as to harvest
trees. You could get a jeep through it in summer if you were sufficiently
motivated - and the owners of the cabins do, from time to time.  I chose to
turn a blind eye toward the cabins when I walked through it with a GPS and
left it in OSM as a track.

By the way, OSM is telling horrible lies in that area, but I don't really
have access to a lot of the land to survey and correct it.  I know that at
least the porton of Johnson Hollow Road is abandoned, although I've seen
signs that snowmobiles use it. It was never much of a road. It was built to
maintain the ski lift on Bearpen Mountain, back in the day that there still
was one. There are a couple of private properties that it accesses, down
below.

There are a bunch of similar ones that are indistinguishable from tracks if
you don't see that there's a cabin. In
https://www.openstreetmap.org/way/239570596#map=17/42.26227/-74.17145
you'll see address points for 351 and 360 Barnum Road. The trail is gated
off, but the two inholders have keys to the gate. The trail is drivable by
a high-clearance vehicle for a little way in. It's not plowed in winter.
Again, OSM is still terrible there. I never surveyed that screwy TIGER crap
to the north on the Barnum Road Unit (New York City land.)  It's at most a
track now, but it's misaligned.

The point here is that in the dense tree cover, you might not spot a cabin
on aerials, and with private property about, you might not have access to
the land to survey in the field.  When what you have is a pair of wheel
ruts that run out of sight into the woods, what do you tag it? The
hardliners on the track/service/residential distinction basically reply
that if you haven't followed it to the end to know what asset it accesses,
you don't know enough to map it. Is that really what we want?  Frankly, in
that situation, I see 'track' vs 'service' as a distinction without a
difference. It's heading to a 'leaf node' in the road network. Why do you
care whether it's purely agricultural or whether someone might have a cabin
in there, unless the farm, forest or cabin is your destination?


> 3) Re: "For abandoned roads"
> - If by "abandoned" do you mean "disused" or "unusable"? This is somewhat
> confusing. I think it is generally considered incorrect to map
> unusable roads (even for off-road vehicles like ATVs or mountain bikes) as
> highway=track if they no longer exist, say if they are completely overgrown
> with scrub or completely washed out by erosion or landslides. While there
> are many non-existent roads currently in OpenStreetMap, that's because the
> Tiger data was often very old, or people have imported old forestry or
> agricultural tracks which no longer exist.
>

Commoner around here is that the agricultural or forestry tracks, or
service roads, do exist, but may be miles from where the census taker
sketched them. It's pretty obvious in most of the rural counties in New
York that the minor roads in TIGER were created by diligently digitizing
pencil sketches drawn by census field workers without any cartography
training.  When doing TIGER matching, it's best to think in terms of, "this
service road turns right off the main road and doglegs to the left - is
there anything anywhere nearby with about that shape?" I usually spot one.

We also have a few examples, such as the aforementioned County Route 3,
where the highway department has abandoned a road that is still in use as a
track, service way, or hiking trail. Sometimes the public retains right of
access, sometimes it reverts to the landowner - the law in the area is
complicated. I've hiked on public roads whose abandonment predates the
automobile.

>
> 4) Re: "rural... away from population centers" - This is somewhat
> misleading. The tag highway=track is often used in semi-developed parks and
> natural areas located right next to cities.  Searching overpass-turbo finds
> over 400 tracks in New York City, mostly in recreational lands, cemeteries
> or parks: https://overpass-turbo.eu/s/141T
>

Some of the ones in parks, and virtually all the ones in cemeteries, are
surely 'service', except that a lot of urban US mappers have the idea
'unpaved=track'.  The ones on recreational lands that are open only to park
service vehicles and don't really go anywhere?  Not sure.  In some cases
they are indeed for urban forestry.
https://en.wikipedia.org/wiki/Urban_forestry But for most of these, you'd
get very little pushback from anyone if you were to retag them as 'service'.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20210223/a46bda46/attachment.htm>


More information about the Talk-us mailing list