[Tagging] [OSM-talk] Should we map things that do not exist?

[Paul Allen]

On Mon, 8 Jun 2020 at 15:40, Mateusz Konieczny via Tagging <
tagging at openstreetmap.org> wrote:

>
> Jun 8, 2020, 16:11 by pla16021 at gmail.com:
>
>
> There may be other indicators of older lines:
> https://commons.wikimedia.org/wiki/File:Stone_sleepers_at_Bugsworth_basin_-_geograph.org.uk_-_450090.jpg
>
> Not sure whatever I am understanding description correctly - it is part
> of railway infrastructure for narrow gauge rail pulled by horses, right?
>

I'm not sure if I understand all of the description, either.  Stone sleepers
like that are typical of a narrow-gauge rail where horses or ponies were the
motive power: wooden sleepers would have presented an uneven surface
for them to walk on, so they walked between stones instead.  Usually
such railways were put in place for hauling minerals from workings, but
might have later been adapted for other uses.

> Also, it is common in my part of the world for older roads, bridleways,
> railways, some farm tracks and even some footpaths to have tree-lined
> hedges.  They're obvious from aerial imagery, although it may not always
> be apparent what type of way they enclose.
>
> If anyone finds image on Wikimedia Commons - please link it here (or just
> add it to a wiki)
>

If somebody wants to pay my passage on SpaceX and can loan me a P900, I'll
try to take a photo for you. :)  Alternatively, use an editor to look at
https://www.openstreetmap.org/way/31982354#map=16/52.0511/-4.6189
especially the section south of Llwyncelyn (look at both the Bing and Maxar
imagery. one gives distinct tree-lined hedges and the other shows the
surface of
the way).

Other indications of former railways are bridges.  There is a typical
style for older railway bridges in the UK that makes them recognisable
as such: https://www.geograph.org.uk/photo/3057076
The line that bridge once served is now the Western Approach Road in
Edinburgh and the bridge is here:
https://www.openstreetmap.org/?mlat=55.94094&mlon=-3.22382#map=19/55.94094/-3.22382
The first time you see what is obviously a railway bridge with road traffic
on it is
a little disconcerting (well, it was for me).

> What, in principle, are the differences between historic maps, a website
> documenting that a route has been constructed over an old railway line
>
> and a sign at the start of the route saying that it follows the path of an
> old railway line?
>
> Is it a sole indicator that route follows former railway line?
>
> Then in all cases I think it is a case of
> "feature is so gone/degraded that it is no longer identifiable based on
> survey,
> requires import of external data to identify it"
>

But aerial imagery and mapillary imagery are external data.  Armchair
mapping
is not based on a survey either.  As I understand it, the very strict
requirement
of only surveys being permitted in the early days of OSM was to prevent
people importing copyright data.  Saying that it had to be based on a survey
was simpler for some people to understand than explaining copyright to
people determined not to understand copyright (like the guy here a few
months ago determined to use copyright data to map watercourses).

>
> I would map it as a property of route (follows_former_railway=yes or
> something)
> if I would want to map that,
>

The only issue I would have with that is the impact on OpenRailwayMap.  I'm
not a railway enthusiast and I've never been a train spotter, but they're a
community that makes use of that data.  I suspect they're also a community
who do the most to maintain the active rail network in OSM,  Do we gain
more by removing a very small amount of unnecessary information, or
chaning how it is tagged, than we lose by annoying them enough that they
lose interest in mapping or move to a different solution?  I know that's not
a good argument for various reasons, but it is something we should bear
in mind.

  cycleway route is verifiable, but route took by army is not)
>

Quite a few motor roads in the UK follow those "unverifiable" routes.  Some
are even named after those routes.
https://en.wikipedia.org/wiki/Watling_Street

I think that for "Here railway is gone without any clearly identifiable
> trace in
> terrain." case it is OK to have  "should not be mapped and can be deleted
> if
>
mapped." ("can be deleted" was just changed from "should be")
>

That's an improvement.  I'm tempted to say that it should suggest contacting
the original mapper before deleting, as the original mapper may be able to
provide evidence of identifiable trace.  But I'm also tempted not to,
because
I know that such a conversation is likely to degenerate into "I want it"
versus
"You're not having it."

Also - it is OK to have
> https://wiki.openstreetmap.org/wiki/File:Galeria_Kazimierz.JPG
> example with
> "Location of a former railway and railway station without any
> traces whatsoever. Not mappable."
> ?
>

It's back to what you consider traces.  I wouldn't map a station building
which
has been demolished and built over.  OTOH, the razed railway that went past
it
aids understanding why the address of the building is "Station Yard" and
it's
on "Station Road."

-- 
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200608/544a3449/attachment-0001.htm>