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

Mateusz Konieczny matkoniecz at tutanota.com
Mon Jun 8 13:41:30 UTC 2020




Jun 8, 2020, 14:57 by voschix at gmail.com:

> Warin, Jack,
>
> your comments are really off my main point.
> We have an unfinished mailing-list thread where we have different opinions on whether a razed (on the ground) railway can be mapped in OSM. 
>
This discussion appeared multiple times and is clear controversial point

> In the middle of that discussion the abandoned railway wiku page gets completely rewritten by one of the participants in the thread explicitly stating that razed railways should be > removed>  from OSM. 
> This is basically against good practice in OSM.
>
It was not completely rewritten.

It used to contain "Most mappers follow a basic Good practice#Map what's on the ground principle, and would therefore regard the mapping of non-existant railways as incorrect data to be placing in the OpenStreetMap database."
https://wiki.openstreetmap.org/w/index.php?title=Demolished_Railway&oldid=1625129


> In addition the statement that where roads trace razed/dismantled railways, the reference to the fact that they do, should be removed is clearly wrong.
>
Given long discussion(s) on exactly this topic "is clearly wrong" seems to be going quite far.

> Worldwide there are many thousands of km of roads and cycle routes that retrace exactly former railway lines . what is wrong with adding railway=dismantled (orrazed)  to the ways that make up the road or the cycle route. 
>
Because OSM is not for mapping completely gone historical objects, this is not verifiable
and it invites mapping further historical and even harder to verify objects and promotes
mapping razed railways where even such "trace" is gone.

And it makes harder to edit actually existing objects, confuses mappers (especially newbies).

(to be explicit: this is my opinion, I know that consensus position is more balanced)

> many country paths and tracks labeled with their Roman or Saxon names, even though the present-day structure is much younger - they only retrace the Roman way like the present-day street in the first example on the wiki page retraces a former railway..
>
There is 0 opposition to mapping cycleways following course of former railway. Sole problematic
part is whatever old railway can be mapped and if yes - how.

> BTW I am not saying that OSM map data are incomplete without mapping old raylways, I am only asking to not remove those that are mapped, and to not write in the wiki that they should be removed.
>
Old existing railways - there is no problem with mapping them
Old railways removed, but with traces remaining - mapping is OK (the tricky part is 
what counts as remains)
Fully completely and totally removed railways/roads/buildings can and should be deleted
and it is improving OpenStreetMap. 

To repeat: the "100% gone object should be deleted" has clear consensus.

The discussion is whatever razed railway can remain mappable - and in many cases it is.
And what counts as mappable railway. 

> BTW 2: wiki pages in general should not invite mappers to remove already mapped objects, but only correct mapping errors. 
>
Removing already mapped objects in many cases is correcting mapping errors or updating map data.

I routinely delete nonexisting shops, nonexisting buildings, nonexisting trees/forests, nonexisting routes,
nonexisting ref, nonexisting paths, nonexisting, roads, nonexisting tracks 
and yes, also nonexisting railways (in my region last one is typically happening with short and minor 
spurs that used to serve industrial areas, now often disappearing as area is rebuilt what erases all traces).

Such edits are improving OSM.

(and yes: where possible I retag unsigned_ref or update to current state like natural=tree_stump or
use demolished:building where visible on aerial images and delete nonexisting railways in cases
only where no identifiable traces remain)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200608/1bc33030/attachment.htm>


More information about the Tagging mailing list