[OSM-talk] stop deleting abandoned railroads

moltonel 3x Combo moltonel at gmail.com
Thu Aug 20 11:55:57 UTC 2015


On 20/08/2015, Russ Nelson <nelson at crynwr.com> wrote:
> moltonel 3x Combo writes:
>  > The demolished: prefix only makes sense when there is something left
>  > of the former feature, typically rubble (useful for example to alert
>  > boattripers of the hazard). When there is nothing left in reality,
>  > there should be nothing left in OSM.
>
> Question: should we tag the aqueduct underneath Sunrise Highway
> between Aqueduct Raceway and Freeport, NY?

I'm not at all familliar with that area, please provide some links.

>  > Deleting an object is hardly different from editing it as far as
>  > osm history is concerned.
>
> Except that deletion excises it from the database that you see when
> make an API call.

So does editing. When you change the geometry or tags of an object,
the old versions are not downloaded/displayed by you editor unless you
take special action. I know that outside of Potlach1 that "special
action" is a bit more complicated, but that is just an API issue that
will hopefully get fixed someday.

> In the case of dismantled railways, that is not
> accurate. There *is* a dismantled railway there, and you can tell
> because the railway was at point A and at point B, and you can still
> see it there, and so you should expect to see it in-between.

The argument (which is not making any progress so this might be my
last comment on it) is between *is* and *was*, and where to draw the
line. If there *is* an abandoned railway it can be mapped. If there
*was* a railway it cannot be mapped. "abandoned" isn't a synonym of
"was".

See for example http://osm.org/go/esz3FWUuB- (toggle satellite
imagery). There *is* an abandoned railway south of the river, there
*was* a railway north of it. The fact that you can infer that the
railway was indeed there because it's clearly visible again at
http://osm.org/go/esz18LcmF- (and visible all the way in the GSGS 3906
imagery) doesn't matter.

We've discussed a few criterias to distinguish between *is* and *was*
on this thread, but you've dismissed even the most basic "A building
has not been constructed at that location" one.

On the subject of is/was criterias, I'd like to weight against the
less basic "railway grade slope" one. Firstly because railways usually
followed existing flat grades instead of following them, secondly
because in other cases the cuttings and embankments should be mapped
for themselves rather than implied by a railway=abandoned. There might
be some cases where that argument still makes sense (montainside
railways come to mind), but it needs to be evaluated case by case
IMHO.

> I understand that most people don't give a crap about map feature X,
> Y, and Z. I get it, really I do. I look at things in OSM myself and
> wonder "why the hell did you map that?? Who cares??" And when it comes
> to railways, there's a lot of people who don't give a crap. Fine. Go
> ahead. Don't care. But I do. So don't delete the things that I (and
> other railfans) have added.

For the last time, this isn't about esoteric mapping topics (abandoned
railways is actually quite popular in OSM), but about reconising then
something just doesn't exist anymore and (in another part of this
thread) about wether mapping the past is acceptable in OSM at all.

> From whence comes this impulse to destroy other people's work? Cuz it
> seems pretty anti-community, anti-mapper, and anti-OSM.

Quality assurance. We all want the map to be as correct as possible,
and that sometimes require deleting data. The only anti-* case is when
the decision to modify/delete is controversial but not discussed.



More information about the talk mailing list