[Tagging] roads with many names
joseph.eisenberg at gmail.com
Sun Aug 18 21:48:09 UTC 2019
Slightly off-topic: one of the posts mentioned that this road is tagged
However, if the road is the main access to a house, it should be
highway=unclassified. Or it could even be highway=tertiary since it is a
country road, if it connects to a number of houses or a hamlet or
place=village along the way.
But please add surface=unpaved or a more specific value.
On Mon, Aug 19, 2019 at 4:49 AM Kevin Kenny <kevin.b.kenny at gmail.com> wrote:
> On Sun, Aug 18, 2019 at 12:28 PM Richard Fairhurst <richard at systemed.net>
> > Rob Savoye wrote:
> > > Where I live in rural Colorado, many of the roads have 3 names.
> > > The county designated one like "CR 2", but often have an alternate
> > > name everyone uses like "Corkscrew Gulch Road", and then many
> > > have a US Forest Service designation like "FS 729.2B".
> > name=Corkscrew Gulch Road
> > ref=CR 2
> > usfs:ref=FS 729.2B
> > I think this holds even if the "county-designated name" is CR 2. The
> > everyone uses" tallies with OSM standard practice; the official
> reference is
> > what we have the ref tag for.
> If there are overlaid refs (and even if there aren't) it's better
> North American practice to create route relations for the numbered
> routes. That also lets the 'ref' be associated with a 'network', so
> that a more sophisticated renderer can get the shields right. Note
> that 'CR' is NOT enough information for a shield - if you look at
> you'll see that Bergen County 3/89 have a style of shield distinct
> from Passaic County 677.
> Simple examples of concurrencies at
> where Court Street is US 9 and NY 9N through the village, or
> where I-890 is concurrent with NY 7 for a couple of exits.
> Obviouisly, arbitrary combinations of network are possible, as at
> where NY 206 shares its roadway with first Sullivan County 179, then
> Sullivan County 91, then Delaware County 7, or
> where US 202 is concurrent with US 1A and Maine 9 on Main Street, and
> then concurrent with I-395 and Maine 15 on the freeway.
> It's best to avoid CR as the network in favour of a string like
> US:NY:Delaware. It's a bad assumption to guess that a renderer could
> deduce this from the borders of the administrative regions. Nearly
> the whole of NY 120A is in Connecticut
> Having the route available also helps sort out the directions in the
> signage, as where the eastbound direction of the bridge at
> is NY 30 North and NY 145 South.
> The recommended network for USFS numbered routes is US:NFSR:(name of
> the forest):NFH for a National Forest Highway, or US:NFSR:(name of the
> forest):NFR for a National Forest Road. If the renderer that I've been
> linking to in this message gets one of those (or just USFS for the
> network, but the same numbers can be reused many times, so it's better
> to have the forest included), it'll produce a shield that looks like
> If a road has no name other than its ref, practices differ between
> putting in noname=yes, just leaving the name off, and giving it a name
> like 'State Highway 9'. I tend to prefer the last of these, since a
> renderer that doesn't have any fallback for ways without a name will
> still at least render something. I also don't disturb 'ref=*' on the
> way when I'm creating a route relation, but when doing rendering of
> shaped shields, I ignore 'ref=*' entirely on ways that participate in
> route relations.
> If possible, (borrowing from another thread), keep the ways in a route
> relation sorted. (If you or your editor can't do that, I'll cope, but
> it makes matters easier for data consumers.)
> These practices are seldom necessary in Europe, where route
> concurrencies are rare, and pictorial markers are not important. They
> are controversial for the rendering of the main map, because the
> requirement is specific to a single locale, and because it is foreseen
> to have an adverse impact on server operations when scaled up to
> planetary size with minutely updates. I work on the issue of route
> rendering off and on, but it's a hobby task so my progress is
> agonizingly slow.
> Tagging mailing list
> Tagging at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging