[OSM-talk] razed railways and other things that don't exist today

[Minh Nguyen]

Vào lúc 16:41 2022-10-26, Zeke Farwell đã viết:
> On Wed, Oct 26, 2022 at 1:11 PM Greg Troxel <gdt at lexort.com 
> <mailto:gdt at lexort.com>> wrote:
> 
> 
>     I think people should keep in mind that a culture of deltionism is
>     demoralizing to contributors and harms OSM more than a few  marginal
>     items in the database.
> 
> 
> This is a fair point, but given how often this comes up, it doesn't seem 
> like it's just a few marginal items.  Also it's just as demoralizing for 
> a well intentioned mapper who maps an area and removes some former 
> railway in the process to then get berated for it.
> 
>     I also agree with stevea@ -- old railways are usually visible in the
>     landscape, and the data about where they were in between visible places
>     seems more useful than harmful.
> 
> 
> There is no question that the data about the location of former railways 
> is useful.  However, data being useful is not the standard for inclusion 
> in OSM.  We map the world as it exists today and this excludes plenty of 
> useful data.  Former features that no longer exist are simply out of 
> scope for OSM <https://wiki.openstreetmap.org/wiki/Nonexistent_features>.
> 
>     Also note that people who do not like railroads often do not see the
>     evidence as well as people who are used to looking for it.
> 
> 
> I support mapping old rail beds as railway=razed where they are visible 
> in forests, fields, and other open land.  These traces are often not 
> visible to those with an untrained eye and that's certainly an issue.  
> However, I draw the line at sections going through buildings, highways, 
> excavated areas, or under water where there really are no visible traces 
> by any reasonable standard.  In these situations, a person with a 
> trained eye may see clues and patterns leading them to the deduction 
> that a railway used to be there, but this is not the same as visible 
> remnants.  This is mapping something that is really no longer there in 
> any meaningful way.

I've deduced underground pipelines by similar methods, for example a 
natural gas pipeline that obviously follows a road as it crosses 
multiple streams above the ground. But I've done so with confidence that 
the visible portions must be connected somehow. There's much less reason 
to assume that the traces of a former railway continue to be connected 
below newer development. The pipeline would also be marked at regular 
intervals, so there's a strong possibility that a field mapper could 
improve upon the geometry that I've mapped from an armchair.

One could describe footway=crossing crossing=unmarked ways as another 
kind of deduced feature, connecting sidewalks on either side of an 
intersection. But there's a practical justification for their inclusion 
(routability), as well as a legal justification in some regions.

Some historic railway mappers would quibble at the notion that they're 
merely connecting the dots. Over the years, I've heard some rather 
tortured arguments, forcefully delivered, about these ways being the 
result of intense surveying. I tip my hat to anyone willing to devote 
their time to finding mappable railroad spikes via metal detector. 
However, we should expect them to document their findings meticulously, 
beyond just tagging source=survey on a railway=* and expecting to be 
consulted when it comes up for deletion many years later.

Personally, I never got into abandoned/razed railway mapping until I 
started contributing to OpenHistoricalMap. Former railroads are 
impossible to ignore when detailing the histories of so many towns 
across the U.S. Indeed, their remnants are all over the place; it's fun 
to realize that a tree line or embankment that you always took for 
granted fits into a hidden puzzle of former railroads. As a layperson, 
I'm only able to make these connections because of historic maps, 
photos, and timetables, the kinds of sources that are irrelevant to OSM 
but central to OHM.

I think historic railroad mappers who limit themselves to OSM's 
railway=abandoned/razed tags are missing out. Why not map the whole rail 
network as it was, unapologetically? In the areas where I map, OHM 
already has more former railroads than OSM will ever have, even despite 
the real ergonomic issues mentioned earlier in this thread.

-- 
minh at nguyen.cincinnati.oh.us