[Tagging] RFC - Discourage railway=preserved

Stefan Keller sfkeller at gmail.com
Sun Apr 11 21:24:52 UTC 2021

This seems to be a rather half-baked proposal. And it could be easily enhanced.

The proposed solution "railway:preserved=yes" is a typical case of
[namespacing and Prefix fooling] (see there). Plus: What is the
meaning of "railway:preserved=no|both"? I see the issue of adding more

So, if you want to differentiate construction aspects like
"narrow-gauge" create a secondary tag (similar to cuisine which is a
secondary tag to restaurants).


[1] https://wiki.openstreetmap.org/wiki/Talk:Namespace#Over-namespacing_and_Prefix-fooling

Am Mi., 7. Apr. 2021 um 17:28 Uhr schrieb Kai Michael Poppe - OSM
<osm at poppe.dev>:
> Hey all,
> In case you didn't see, I created a first experiment to fix that with just trying to create parity between both renderings - and I my test with the latest Geofabrik-PBF from DE-NDS (using hstore, no problems on that end), it worked ok (apart from the name not showing correctly): https://github.com/gravitystorm/openstreetmap-carto/pull/4361 - An example with old and new is shown.
> Kai
> On 07.04.2021 13:48, Jeroen Hoek wrote:
> > On 07-04-2021 12:14, Mateusz Konieczny via Tagging wrote:
> >> Note "Database layout change" part - right now this tag is not in a database
> >> used for rendering :(
> >>
> >> So it may require changing in database import definition and deployment will
> >> be definitely delayed and/or more complicated.
> >>
> >> Though it is also possible that OSM Carto has no need for yet another barely
> >> distinguishable and unclear rendering difference, maybe railway=preserved
> >> special rendering should be dropped?
> > As Marc_marc also noted, I think you can just use the hstore tags column
> > to filter on those (I know this works in my airfield rendering
> > experiment), unless that is not an option because of performance. Do you
> > know who we could ping in that Carto issue to answer that question?
> >
> > Dropping rendering is a choice that the Carto maintainers can make of
> > course. I personally don't have a strong opinion either way (slightly
> > preferring some form of differing rendering), just that rendering the
> > old method but not the new method is undesirable at this point where the
> > new method has superseded the old one.
> >
> > _______________________________________________
> > Tagging mailing list
> > Tagging at openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/tagging
> _______________________________________________
> Tagging mailing list
> Tagging at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging

More information about the Tagging mailing list