[OSM-talk] stop deleting abandoned railroads

moltonel moltonel at gmail.com
Sat Aug 15 22:19:52 UTC 2015

On 15 August 2015 16:29:56 GMT+01:00, Russ Nelson <nelson at crynwr.com> wrote:
>Are we even talking about the same thing? Let's assume that you made a
>s mple t po.
>Don't those last two words look a little weird with missing bits?
>Shouldn't those letters be there? Shouldn't the dismantled bits of a
>railroad be in OSM as dismantled bits?

That metaphor is a bit of a strech. Let's bring it closer to reality with http://osm.org/go/esT4qhWnF these WWII signs which are technically just stones in a field. What if some real-life vandal removed the stones of one letter and nobody repaired the damage long enough that all signs of the former letter are gone ? You're arguing for some kind of *=razed tag, while I (and probably most other osm contributors, who arent wwii-stone-markings enthusiasts experts in deducing a former sign from peripheral hints) would argue for deleting the non-existing letter from osm.

> even if it's only because you can look left
>and see evidence of the railroad, and you can look right and see
>evidence of the railroad. Should they NOT be mapped through the
>farmer's field where they have been plowed into dismantlement?

It's a circular argument at this stage, but yes if the ground there has been flattened and ploughed, osm should IMHO not map anything else than the field. I'd support deletion of that railway section in such a case, but of course it should be discussed with the other contributors. As a last resort, the DWG can arbitrate between two parties.

>Now, I'm sure somebody will, at some point say, "Russell, just go off
>to OpenHistoricalMap and put your data there." That's fine, except for
>those pesky implementation details where THEY ARE IN TWO DISPARATE
>DATABASES, UNCONNECTED. How, exactly, do you make a relation that
>shows the entire route of a railroad when half of it is off in a
>different corner?

That's clearly a big pain point of ohm. But it isn't an insurmountable one, hopefully ohm will eventually manage a continous merge of osm data as a 'present day layer'.

>I don't understand why we're having this argument. We map tons of
>things that you can't see. Why not map as dismantled railroads that
>have been dismantled? Why not make an exception to the "Delete it if
>you don't see it" guideline?

The existence of ohm is a strong aknowlegement that osm is only for the present. Russ, you're an expert in old railroads, but think of all the other old things you could be an expert of. If all the niche experts got their exception, the osm tools, cpnsumers, and contributors would suffer heavily from all the historical data. In its curent state, osm isn't fit for historical data (end_date and other lifecycle tags are only good enough for some narrow cases of objects that still exist in some deteriorated form and haven't been recycled yet).

Hopefully someday the ohm framework will be mature enough to be adopted by osm, so that we can map in time as well as in space (better tools to map in the 3rd dimention would be great too). In the meantime, please only map the present in osm.

Vincent Dp

More information about the talk mailing list