Proposing to deprecate railway=razed and railway=dismantled

no, an abandoned:railway is not usable as a railway whatsoever, compared to a disused:railway which could be used but isn’t.
dismantled:railway or razed:railway have been actively removed (but not all traces are removed completely, or they wouldn’t be taggable).

For buildings, as long as they are buildings, I’d rather use abandoned=yes or disused=yes, and not the lifecycle prefixes, similar to a quarry.

If walls and roof are gone, it surely isn’t an abandoned building any more, but ruins or lesser remains.

Yes, we map that there IS what is left from a railway. We map what is observable.
Here’s a well documented example: Bahnstrecke Berlin-Wannsee–Stahnsdorf – Wikipedia

I think this is not the interesting question for the taggability discussion, it is not about how likely there could again be a railway, it is whether the impact the thing had when it was working, is still perceivable today.

5 Likes

hard to be 100% sure but it looks like cutting remained here, and may be more defensible than vast majority of railway=abandoned

And I’m not saying that abandoned railways aren’t mappable as railway infrastructure either, I’m just saying that to call this (e.g.) “abandoned:railway” :

is, in my not-so-humble opinion, a fantastical disregarding of any and all common sense. To me, it must still be a railway in order to be considered an abandoned railway. And that, above, is not an abandoned railway: that’s a gravel road.

I like your bus stop example. I’m not arguing ‘highway=bus_stop’ is akin to calling a bus stop a “highway”, what I’m arguing is that when a bus ceases to stop there, it ceases to be a “bus stop”. Can I tell that a bench on a concrete pad immediately adjacent to a street was, at some point in the past, a bus stop? Sure. But unless buses stop there and let people on and off: it isn’t a bus stop anymore. It has lost the intrinsic, fundamental characteristic that made it a bus stop. Now it’s just a bench on a concrete pad.

Don’t get me wrong, I can imagine scenarios where keeping a long-since-abandoned piece of infrastructure on the map is the most accurate thing to do. Dieter linked to a Wikipedia article above:

I look at the photos on that Wikipedia article page and I say to myself “YES! THAT’s an abandoned railway!”:




Where I’m having a big, big problem with the way the tagging is being used is it’s mapping things that are clearly not extant anymore, e.g. the farmer’s field I linked earlier. The “Overton window” of acceptability for these tags has been pushed out so far that it’s reached the point it has exceeded reasonable believability. I don’t know precisely what “the old Harcourt Street line” looked like in its derelict state, but you wrote

The fact that the alignment was chosen to build a tram route, in preference to any other possibility, shows that it was still a meaningful piece of physical infrastructure, not just an idea that “something used to be here”.

And I strongly disagree, because you also wrote:

It was closed in 1958, […] But most of the alignment remained and was protected from construction.

Sounds to me like the alignment was chosen to build a tram route because there was an absence of infrastructure.

That’s where I look at that photo from the examples on the Key:abandoned wiki page:


And I’m just left shaking my head at the argument that that’s “railway infrastructure”. Mateusz noted above that:

But a “cutting” is not “railway”. A cutting is just the end-result of the removal of a pile of dirt. A railway that was built along a cutting does not mean that cutting is forever more a “piece of railway infrastructure”: it stops being “a piece of railway infrastructure” when the railway no longer exists. From that point onward it’s just a cutting through a hill. It is not “infrastructure”: it is merely the absence of dirt from where it used to be.

Here’s a local-to-me example where fortunately no one has been so bold as to run an abandoned:railway=rail up it:

You see that space between the chain-link fences? That was a siding, at one point. It used to do this:



That alignment, very roughly, is going to be re-used by a new LRT line. In point of fact someone added the proposed new line to OSM three years ago: Way History: ‪Green Line Phase 1‬ (‪860275662‬) | OpenStreetMap

And as you can see, a certain someone changed “construction” tags to “proposed”, because it was going to be a number of years before shovels hit the ground… :upside_down_face:

Should there be abandoned:railway=rail there? disused? razed? Because that’s the argument that’s been made to me, as far as I can tell. As as per the previous comments that is… hmm, well, I’m running out of synonyms for “insane”; I may need to consult a thesaurus.

3 Likes

I think here is crux of misunderstanding. You seem to be under misguided impression that dictionary definition of English word “railway” is the same as all OSM tags which include ASCII text “railway” somewhere in the key name. (I might be oversimplifying a little here, but hopefully you’ll catch my drift).

That is hugely incorrect assumption (and not just in railway case, but in many OSM tags).

Even basic (non-prefixed) OSM railway=* tag contains way more things than just the “metal rails” (e.g. railway stations, railway signalization, passenger platforms, ventilation shafts, water cranes, roundhouse service buildings, buffer stops, derailing devices, transfer tables and many more).
They all have something to do with railway infrastructure, though, which is vastly more diverse term then “railway rails”, which you seem to insist on as only possible righteous combination of words.

Even more so, different tags which happen to contain word “railway” in the name (like e.g. abandoned:railway=*) have even more diverging meaning. As noted several times by several different people, abandoned:railway=yes is not railway=rail. Not even close. Just like your ruins:building=yes is not a building=yes, not by any stretch of imagination. If it were similar, it would use same tag (and then add some additional tag for differentiating details; e.g. amenity=restaurant + cuisine=*).

So, to summarize, things which you seem to consider “railway” (i.e. railway=rail) are just a small percentage of things (quick look in taginfo would say about 10-20%?) of the things which are mapped in OSM and contain word “railway” in tag name.

In other words, railway=* tag is not about Railways. It is about railway infrastructure, and were it not for historical accident (and people preferring shortness), should’ve been more properly called railway_infrastructure=*. Would you agree? And, if you do agree, and can try to imagine the tag is indeed called railway_infrastructure=*, would abandoned:railway_infrastructure=yes make more sense to you in cases where just parts of railway infrastructure remain?


rails + ties + ballast + signals + switches + catenary masts & wires = railway

Also, I’m wondering how would you tag if there were ties + ballast + signals + switches + catenary masts & wires but only missing thing was metal rails (i.e. situation which according to you “is not a railway”)?


As you say you’ve done ground survey there, I will take your word for it! And just to be clear, as you say it has been just a farmland with no trace of railway for almost 5 decades, I will take your word for it too, and in such case I definitely would not tag it abandoned:railway=*.

You want abandoned:railway=rail to be there so badly that you’re seeing things that aren’t there.

This might came as a surprise, but I don’t really care much about that example :smiley:
Instead, I have two points why I replied to it:

  • imagery is often bad, so people should not go around willy-nilly deleting other people work, unless they’ve surveyed it on the ground. And if they do delete it (because as you say “no visible trace of it remains” and there is no chance that people will re-add it from satellite imagery or similar), they should note in their changeset comment when and how they’ve verified it doesn’t exist any more.

  • For this particular example and screenshot I posted: Imagine, if you will, a railway fan who knows the railroad went there at some point in time, and can see those parallel black streaks at exact position where rails used to be. Pareidolia is a known issue with human brain (I mean, if you don’t see that face on Mars, something is wrong with you :smiling_face:).
    So, I’ll tell you what will happen - that mapper (not me, I don’t map railways which I didn’t survey in person) will likely add new way tagged railway=rail (or similar) on the OSM. Then some other person will delete them, and then the cycle will repeat.

    So, in cases when things which look like a railway are visible on recent imagery, and they pass where railway used to pass, and someone has mapped railway=rail or even railway=abandoned while in reality no traces remain; I would recommend not deleting those ways, but instead tagging them as razed:railway=yes (or railway=razed which seems preferred by some) or similar schema (e.g. in this case, if you say it’s a dirt track, you might use highway=track + surface=ground + not:railway=rail) and document in changeset that you “verified by on-the-ground survey on date such and such that no traces whatsoever of railway remain, and that the dark streaks that are visible on Bing are instead tractor ruts in the soft earth”. That way, with just a little effort, a lot of anguish can be prevented and situation clearly documented for anyone who later stumbles upon the same issue and mapper time and sanity saved. And at the same time map will be correct and data consumers won’t be confused. All win.

    And when common public satellite imagery gets much better in a few years so it is clearly recognizable to anyone that this is not railway (nor remains of one), only then delete it fully (but also with changeset comment explaining it).

5 Likes

Besides the hair-splitting over tag design, which is nothing new for OSM, you’ve highlighted two points that I think get to the heart of the debate.

First, anyone who can confidently extrapolate a whole miles-long railway line solely based on a metal hook embedded in a gravel path has my respect. I nearly flunked my university archaeology class for failing to make similarly grand inferences from shreds of observational data. It just never clicked for me, but evidently there is an art and science to it. If we frame the activity of abandoned railway mapping as a kind of industrial archaeology, maybe it would find more respect within the community.

Second, who has the burden of proof regarding rail infrastructure in case of doubt? The rail mapper who asserts that some relic of the infrastructure still exists in place? Or the layperson who is skeptical of the claim and unable to verify it despite doing due diligence? Whose starting assumption best serves the end user of OSM data? Does it matter when a local mapper feels outnumbered by remote mappers?

Technically, the on-the-ground rule means we aren’t supposed to rely solely on published documents, let alone historical ones. But what if the rail mapper is surreptitiously relying on those documents and the rest is post hoc rationalization? That puts the layperson at a serious disadvantage. The rail mapper could claim many untruths and leaps of logic without the kind of scrutiny normally applied to other map features, sending the layperson on a wild goose chase.

To be clear, I believe most rail mappers are trustworthy, upstanding members of this community. I think they care too much about railways to ever dream of smuggling fake data onto the map! But I see a lot of incredulity directed towards them, perhaps because of a few overzealous rail mappers over the years who have given them all a bad name.

The rail mappers can resolve this misunderstanding by more rigorously documenting their sources and methods. After all, not every published source is equally reliable. Not every inference is unimpeachable, immune to reconsideration. We need to be able to have that discussion rather than this tug-of-war between rails=good and rails=bad. In OHM, we have a longstanding practice, not always followed very rigorously, of documenting projects on the wiki where we can elaborate on our process.

When you suggest deleting the railway with documentation of an on-the-ground survey that rules out alternative explanations, it strikes me as a much higher standard than the rail mapper applied in the first place. It’s almost too long to fit in a changeset comment. :wink: Is that fair because it’s much harder to create knowledge than erase it? Is that fair because it takes a lot of time and effort to upgrade our crude contributions from a decade ago to today’s more rigorous norms, or to migrate them to OHM? OSM is never in much of a hurry anyways.

4 Likes

You’re “oversimplifying” to the point of being condescending and insulting. I understand precisely what you’re saying, that railway within the context of OSM tagging is not “railway” within the context of the dictionary definition. Disabuse yourself of the notion that I don’t understand your position: I just think it’s preposterous.

All your jibber-jabber about how railway=* encompasses more than just railway=rail is a diversionary tactic to distract away from the underlying issue: overzealous people mapping razed:railway=rail, disused:railway=rail and abandoned:railway=rail where there is no railway=rail.

Where it’s really bothering me, going back to my wonderful example of the non-existent railway in the farmer’s field, is that there seems to be a very enthusiastic group of mappers who are giving primacy to railway tagging on patches of dirt that once had a railway on them. Again: the “farmer’s field” example had a way tagged highway=path, which was later concurrently tagged railway=abandoned, until someone in the name of “cleaning up” the so-called “misuse” of tags deleted the original highway=path. That is categorically incorrect. It’s a blatant falsehood.

Well a rail switch without rails is non-existent, so that’s moot. A catenary mast is just a very specific type of power=pole, it’s wires just a minor power line. Of course without the rails there are no trains to power, so I wouldn’t map these at all because they’d be de-energized anyway. Without the rails a railway signal is entirely pointless and non-functional, so likewise I wouldn’t map them; without the rails there are no trains and without the trains those signals will never ever light up again. They’re just never-illuminated lamps on a pole. And as I wrote before: ties and ballast without rails on them are just pieces of wood and gravel.

I wouldn’t tag anything; I wouldn’t map any of it.

Century. Half a century. That is how absurd that is.

1 Like


Preach, brother, preach! Can I get a “hallelujah”?!

But we don’t only map things because they are currently useful for their original purpose. If that was the criterion we would need to remove most ruins and historic and archeological sites from the map. Mapping to represent the physical environment is also legitimate. To give one example, abandoned railways and ruined buildings are very useful for orientation when hiking.

Again: I am not defending the farmer’s field example where none of this seems to apply.

11 Likes

This bothers many of us. However there are also plenty of places where terrain features of former railways are observable, mappable landmarks. I wish that tagging described the current terrain feature rather than the abandonment of the former feature, but changing it at this point is probably not realistic. Richard explains the legitimate uses of these tags well up thread:

3 Likes

I apologize if you understood it as condescending, (as stated) that was not my intention. But as you simply ignored several previous attempts to get an understanding on that critical point, I felt I had to ask harder. Regardless, I am glad to get the confirmation that you agree that railway=* tag is about all railway infrastructure !

Now that we have common ground, I’ll repeat my question which you seem to have missed:


You seem to continue with (if I understand correctly) that you consider it a problem that:

What? This part sounds like oxymoron to me. You seem to be saying that there is a problem when people map e.g. abandoned:railway=yes when there is no railway=rail? Was that a typo? Or, if not, could you clarify?

Of course one would only tag abandoned:railway=rail when there is no railway=rail. Same with razed:railway=*. One should never use those prefixes when there is railway=rail on the ground! If you used razed:railway=rail when there is actual railway=rail on the ground that would always be incorrect.

If there was railway=rail (i.e. operational standard-gauge railroad), you should always tag it as railway=rail. I’ve took that as self-evident all this time, so I am hugely surprised what you meant by that quoted text above.

Again, I get the impression you’re replying to my posts, but arguing with someone else. Did I somewhere give you the impression that removing existing highway=path was valid? Did someone else in this thread? I even explicitly said for that particular example that I completely agree with you that railway=abandoned is incorrect tagging when nothing of railway infrastructure remains, and that I would never tag it abandoned:railway in such cases, i.e.:

Nor have I seen anyone else in this thread claiming otherwise (i.e. nobody claimed here AFAICT that abandoned:railway=yes should be used when no traces of railway infrastructure whatsoever remain on the ground”). Did you see it and I missed it? Could you link to it? Or perhaps you have seen such discussions elsewhere (outside of this thread)?
I mean, I see little point on getting so riled up on a specific point, in a thread where everybody seems to agree with that point. :smiling_face:

Thanks, I guessed as much, but had to confirm, lest I assumed incorrectly. Now, the more important followup question: if such remains of a railroad (i.e. everything but a metal rails) were tagged as an OSM way with abandoned:railway=yes, what would you do? Ignore it? Delete way? Retag it with something else?

Of course; lapsus calami, sorry 'bout that. (I’ll correct original post after I finish composing this one).


I’d also be grateful if you’d care to share your opinion on those two points in last paragraph in that post of mine you’ve replied to (i.e. about potentially incorrect taggings and suggested procedure of addressing it). Do you see the issue I am pointing to? Would you agree with some of my suggestions on how to address it? Or, if not, would you propose some alternative solution that would address issues mentioned there in some other way?

2 Likes

I usually employ combination of several principles (not just about railways, but about all OSM data, and even ancillary services like wiki):

  • burden of proof should generally be on person deleting the data. (i.e. I highly prefer avoiding loss of useful data to the cleanup of potentially useless data, a principle I employ in OSM software too)
  • on-the-ground work (especially if documented with pictures) usually trumps armchair mappers (especially with bad imagery or without well documented source in changesets)
  • mapper reputation (quality of their other work, DWG/wiki blocks, OSMCha reputation/warnings etc.)
  • content of changeset discussions regarding disagreements in edits
  • individual manual work vs. mechanical edits
  • all other things being the same, in case of tagging conflict of domain-specific tags, prefer the domain expert opinion to that of layman (i.e. if telecom installation technician claims details of some man_made=street_cabinet, they are more likely to be correct than I as passer by, even I happen to walk by while it was open and they were working on it)

For example, majority of my OSM work on railway infrastructure has in fact being in verifying on the ground situation, updating incorrect railway=rail to abandoned:railway (with wikimedia_commons=* pictures of remaining railway infrastructure) or both of them to razed:railway where I could no longer find any traces of railway, and adding/updating missing pieces of (sometimes non-rails) railway infrastructure - all with detailed changeset comments (as I suggested others also do in my previous post)

(Sure, I’d love to add new railway=rail infrastructure too, but those seems well-mapped in my area and rarely new ones are built, and I prefer on-the-ground mapping)

4 Likes

Hmm, really? I’d say the burden of proof is generally on any person changing any data whether that be an addition, modification, or deletion. If someone adds a suspiciously fictional looking feature without providing a shred of evidence for its existence, we rightly revert (delete) that change. We don’t hold the reverter to a higher standard than the original mapper (a very low bar in such a case). I think what you mean to say is that if a feature was originally added with proof of existence, it should not be deleted unless a similar level of proof is provided that it no longer exists. Yes? This I agree with.

6 Likes

:+1:

These are all common-sense considerations, but they rely on trust – particularly, trust that the other mapper is who they say they are, doing what they say they’re doing. Whenever we see this perennial topic flare up, it’s because, unfortunately, that trust has already broken down somehow.

Sometimes it’s because non-railway mappers lack a full appreciation of the richness of rail history. Other times, a mapper suddenly tells us about a recent field survey they carried out, only after someone calls them out about an abandoned railway they added and threatens to delete it. No harm done, I suppose, unless it becomes a pattern. When a mapper copies from a proprietary map or imports something without following the import guidelines, then tries to pass it off as more organic mapping, the community recognizes that as a problem. We should recognize that Wikipedia and old maps can tempt mappers in the same manner as Google Maps and GIS shapefiles.

Personally, I haven’t bothered fighting the addition of tenuous abandoned railways around me, because there’s enough else for me to do around here. But once in a while, I have had to clean up overzealous historical mapping. I think many general-interest mappers have had similar encounters, conditioning us to be more skeptical of anything related to historical rail data in OSM.

For example, back in October, I had to surgically remove historical information about railway stations along a former passenger line, most of which had been demolished or repurposed in the intervening decades. It was added in a series of tiny edits by a mapper who specializes in mapping rail routes along abandoned railways, representing lines that haven’t run for decades. I only noticed their edits when a station popped up in my hometown even though it closed in 1967. At least that building still exists; for some other stations, they just plopped a node at an arbitrary location beside the tracks. If they did conduct a ground survey, I would love to see a photo of ticket stubs embedded in the ground, proving the existence of this rail infrastructure.

This is an illustration of the slippery slope. As we accommodate the intrepid rail mapper who takes their metal detector to the field to find buried spikes, we also accommodate less determined mappers who, for all we know, are just synthesizing information they read online somewhere. In the grand scheme of things, even false railway=abandoned ways don’t harm the map that much, because most data consumers ignore this tag anyways. But should editors then add presets for abandoned:railway=station so these mappers don’t cause more practical problems by reaching for railway=station? Will this in turn encourage even more tenuous mapping of railway infrastructure, and now of buildings too?

2 Likes

Though I would not expect on-the-ground checking where it is obvious that original import was not based on the on the ground survey (for example if someone imports 10 000 railway lines that turn out to be not existing anymore it is not necessary to check each of them).

Or when edit is described “digitizing 1930 railway map”. Or where author uploaded many utterly gone railways and stopped responding to changeset comments.

Sigh… Matija, we’re just circling the wagons on this. Talking past each other. Like, this entire part of your post:

:man_facepalming:

I don’t know how to break my point down any simpler.

When there is no railway=rail, it has ceased to exist as railway=rail. Thus, to tag a way along the path where a railway=rail used to run—but no longer does because it doesn’t exist anymore—with abandoned:railway=rail, is nonsense.

I understand that abandoned:railway=rail such as it’s defined in the wiki and such as you think it ought to be used says “this is for identifying where a railway=rail used to be but only a portion of it remains”. What this has led to, the way it has been interpreted by many mappers, is that any shred of evidence that a railway ever existed there should be mapped as abandoned. What I keep trying to tell you, over and over and over and over, is that I am fine with the use of abandoned:railway=rail where the railway=rail still exists but very obviously isn’t used, won’t be used, and has been left to deteriorate. But there comes a time in that part of its “lifecycle” where it has deteriorated or been actively demolished such that no reasonable person can say “that’s still a (lifecycle_prefix):railway=rail”, and at that point it’s something else entirely. Could be highway=path, could be something else; could be nothing.

Oh really?

:unamused: :man_facepalming:

Of course, but I meant that in assume good faith (by default i.e. unless shown otherwise)” meaning. E.g. if somebody added a new shop in OSM that I didn’t notice while I was walking there, then my first reaction should not be “let’s delete it ASAP”. But instead: ask the mapper who added it in changeset comments, or go there verify it myself, or at least check a website if they have a shop there, or something similar.

I think what you mean to say is that if a feature was originally added with proof of existence, it should not be deleted unless a similar level of proof is provided that it no longer exists. Yes? This I agree with.

If you replace “similar level of proof” with “at least a tiny little higher level of proof”, yes, that is exactly what I meant. (e.g. in that example above, if one party added a shop and claims they’ve seen it there on the ground, and other party deleted it on the same day and claims they didn’t see it there on the ground on that same day; and everything else being the same, one should leave that shop on the map).

(that might be my programmer inner-self speaking; because if there is if (a=b) then b that iterates for all a,b it will produce infinite loop by making “b” correct in first pass, and then making “a” correct in next pass, ad infinitum. There must be some tie-breaker to trigger if everything else is completely equal, and for me that tie-breaker is “keep data” in precedence to “delete data”).

Of course I agree with you that if the evidence of non-existence is bigger than an evidence of existence, that the right thing is to delete it (e.g. recent vandalism waves as obvious example).
But for disputed ones which are equally likely, there should at least be attempt at changeset discussion before deleting data (again, unless its obvious high-profile mass-vandalism).

Yes, I’ve tried to cover that with that “mapper reputation (quality of their other work, DWG/wiki blocks, OSMCha reputation/warnings etc.)” point as well as other (mechanical edits / imports, changeset discussions). As you say, if some mapper seems to have the pattern of constantly adding suspicious data, that will affect their reputation and might make any new edit they make “suspicious by default” (instead of normal “assume good faith by default”)

OSMCha provides a method to “vote” on changesets and also “follow” users that you’ve determined are suspicious.

I would love to see a photo of ticket stubs embedded in the ground, proving the existence of this rail infrastructure.

That’s why I add wikimedia_commons=* on such places (with all the metadata, including GPS location of the camera, timestamps, etc). Even for places I’ve determined no longer exist I leave StreetComplete notes with pictures showing the area, so people can verify and perhaps correct me.

recent example e.g. just a few days ago I had such a situation, where there is apparently a bakery which was closed with unsigned rolling shutter every time I went by for months; but it is operational, just so happens it only works in the mornings while I'm still asleep :smiling_face:. I'm still working on determining exact opening_hours, but it is definitely closed after noon.

I however am still eating the evidence that it does in fact exists, that was acquired by my informant yesterday.

Of course. I’ve tried to cover the imports by that “mechanical edits” point, as those are suspicious of low quality even in the best of the cases (and majority of imports do not fall under “best case” scenario IMHO). OSMCha does try to tag those as imports for extra validation attention.

Also that partially covered by that “mapper reputation” point: if you have added a ton of invalid stuff on the map in areas I watch, you will get on my watchlist and I will treat all of your other edits as suspicious by default (at least until such time I determine that the mapper has changed their practice for the better).

2 Likes

Sorry, I meant that sardonically, but it didn’t come through. There’s no way a paper train ticket would be recognizable after 40–50 years of sitting in the dirt field next to the railroad right of way. And even if there were, would that really tell us anything about a station building? Occam’s razor says they just picked out an empty lot in a town mentioned on some railfan site, like an armchair real estate agent for urban explorers.

Fortunately, no one has been silly enough to make that specific claim about ticket stubs against my scrutiny of their changeset, but I’ve seen justifications nearly as sense-defying from time to time.

To be clear, there’s nothing wrong with estimating the location of something in the absence of better sources. But a mapper can’t do so while simultaneously claiming that their observations are infallible.

1 Like

I think misunderstanding lies in different definitions.

When you say that “there is no railway=rail, what do you mean by that railway=rail? The wiki definition, or some different definition of your own? If your own, can you clarify exactly what?

Because, for me, “there is no railway=rail means “There is no operational standard-gauge railroad here” – nothing more, nothing less.
Which is AFAICT both the documented usage on wiki, as well as widespread use in OSM database.
(That is BTW exactly why I used tags instead of English words in this – to avoid possible ambiguity meaning of plain English word “railway”, as those OSM tags are much more precisely defined by single source - OSM wiki.)

So while “there is no railway=rail might be true in some situation, there might very well be any of the dozens of other tags indicating presence of some other railway infrastructure, for example:

  • “operational narrow-gauge railroad” (to be tagged with railway=narrow_gauge instead of railway=rail), or
  • “complete, but non-operational standard-gauge railroad” (to be tagged with e.g. disused:railway=rail)
  • “operational tram railway” (to be tagged with e.g. railway=tram)
  • “partially dismantled railway of unknown type” (to be tagged with e.g. abandoned:railway=yes)

Given that, it seems to me as perfectly normal to tag abandoned:railway=yes when there is no railway=rail because it is non-functional due to some (but not all) critical parts of railway infrastructure missing there.
If fact, to me, that is only acceptable use of abandoned:railway=* tag, as documented on that wiki.

If the trains are just not operating currently (even if that situation has been happening for quite some time!), but all parts of railway infrastructure are present, and it can be put back into operation easily, than it is just disused:railway=rail and not abandoned:railway.

AFAICT, that seems to be the whole difference between those two prefixes, as documented on the wiki.


Oh really?
or this (with just wooden sleepers remaining, and all metal long scavenged):File:Dismantled railway looking north 3 - geograph.org.uk - 485512.jpg - Wikimedia Commons

Yes, really. I absolutely find sleepers / ties to be (critical!) part of railway infrastructure.

They are documented as part of railway infrastructure on OSM railway wiki (look under “Terminology” section), on wikipedia, by railroads.dot.gov, by Deutsche Bahn, by Croatian railways infrastructure company (HŽ Infra) and by any other railway-related source that I have ever seen.

ALL of those (hopefully relevant) sources claims sleepers/ties as parts of railway infrastructure, so I’m very curious what is your source for such a claim that sleepers are not a part of railway infrastructure (which you seem to be claiming, if I get you correctly)?


BTW, you seem to again have missed three of questions I reposted for second time, which I think are important for trying to understand each other (and to find common accepted actions for incorrect tags), so I’d appreciate if you would give your view on them!

(or at least reply that you’re intentionally not wanting to answer those if that is the case, so I can stop trying to remind you. At least first two questions could be answered by just a word or two, so I hope that wouldn’t be too much of inconvenience to ask? Thanks for your consideration!)

5 Likes