[Talk-transit] Old railways
jarek at piorkowski.ca
Sun May 12 15:48:14 UTC 2019
On Sun, 12 May 2019 at 07:54, Tijmen Stam <mailinglists at iivq.net> wrote:
> In my environment, some people are adding old ("razed" railways to
> openstreetmak, of which no trace is visible in the field.
> It concerns both old railways which have been gone since 1933, e.g.
> https://www.openstreetmap.org/way/592259029 (Note that this piece still
> is somewhat visible, as it is now a road and partially a cycle path).
The "canonical" answer is that things that no longer exist in real
life and there is no trace of them do not belong in OpenStreetMap.
How strict you want to interpret this probably depends more on local
community consensus than on talk-transit guidance.
Tagging of removed railways that are now paths _in the same alignment_
seems relatively uncontroversial. https://osm.org/way/583243933 is an
example local to me. Your example of way 592259029 seems to me a bit
ambitious in that it traces alignment where it is no longer evident,
such as over houses, and https://osm.org/way/592259043 is a bridge
that no longer exists... I would not include this in OSM.
> Another example is a tram line in Amsterdam that has been gone for a
> year now <https://www.openstreetmap.org/way/672838805>, the area has
> been completely redeveloped, no trace of the old tram tracks remains.
IMO this should not be in OSM.
> I only recently found out about openrailwaymap, but I can't find much
> information about it. It seems it gets its data from the OSM database.
> Is there a way to store "razed" railways somewhere else, so they will
> show up on openrailwaymap but not on OSM (they are rendered on some
> renderers, e.g. OSMAND)
There does exist
https://wiki.openstreetmap.org/wiki/Open_Historical_Map which is a
separate database intended for things that used to exist but don't
Although this would be technically and legally possible, I doubt that
OpenRailwayMap currently integrates data from OHM.
IMHO if they want to show historic railways they should also include
OHM data, to support not having undesirable data in OSM. But I realize
it'd be quite a lot of technical work.
> Funny anecdote: OSMAND showing abandoned railways has on one occasion
> led me to a detour because I thought I saw a footpath cross a canal
> where a razed railway has a very similar rendering.
Arguably this is a bug in OsmAnd - things that are explicitly tagged
as no longer existing should probably not be rendered by a
However this also depends on the tagging being correct - for instance
OSM-carto renders railway=disused (there is a clear sign of a railway
but it is no longer used) but not railway=razed (no railway exists) or
railway=abandoned (tagging seems inconsistent  so I guess they err
on side of not showing disused things).
 P.S. I just realized 3 months ago I tagged railway=abandoned on
stretches where there is no trace of track but its past presence can
be derived from track remaining on either end of the cleared stretch.
I currently find my past decision questionable, but I imagine there
was tagging guidance somewhere that made me choose this over
More information about the Talk-transit