In recent discussions here on the forum I learned, that being narrow is what makes something a path. Should an upper width limit be put into the documentation?
To stay ahead of questions, poll below:
IMHO the relevant question for path vs. track isnât âis it narrow?â, rather, itâs âis it typically used by 4-wheel vehicles?â.
That does imply a track must be wide enough to be used by 4-wheel vehicles. But there are also wide paths that arenât typically used by vehicles. So itâs a necessary but not sufficient condition?
is it a trick question? Or attempt to make ââ (it is not about width only)â only valid answer?
has anyone claimed that?
and answer is clearly not, narrow walls or narrow buildings are not paths
and you can also have narrow footways and cycleways and corridors and so on if you limit to highway= family
also, you can have very wide paths
maybe you meant a different question: is being very narrow disqualifying something from being highway=track ?
Didnât you mix that up? Isnât it sufficient, but not necessary?
And my answer to that question would be: Yes - it canât be a track if the typical agricultural vehicle used in the area doesnât fit. That might be a lot smaller e.g. in orchards compared to vast corn farms.
Funnily, only one person voted for a threshold. In another topic I learned:
From what I observe: A forestry track my be out of use for a dozen years. When vegetation encroaching such a track turns it into a path, then local knowledge means, having seen the location at one single point in time. Just last week I passed over a track that was freshly cleared, which I remembered to be path-narrow two years ago. I did not change the mapping then, because the narrowness was not due to design.
Perhaps. Yet, I still do not see how something once turned path will ever become track again, if there is no upper limit on path width?
I have no idea how you reached that conclusion.
there is some overlap in width between very wide highway=path and narrow highway=track
but âthere is no upper limit on path widthâ does not imply âhighway=path never can become highway=trackâ
there is difference between wide path used solely by hikers and not used by say tractors (in part because they would need to be airlifted there) and highway=track used primarily by farmers to move their equipment
noone is obligated to map, but fixing highway=track overgrown and turned into highway=path by mapping current state would be a correct edit
it does not matter that plant overgrowth was not due to design
So, if I had swapped the track to path two or three or four years ago, I should have changed it back to track this year. I do not think that Iâd do that.
and there is nothing wrong with that
it highly depends on how actively you map, what you prefer to map, mapper density in your area
(but updating it would not be really wrong)
As with many highway type features it comes down to vibes. What one mapper or country might consider a path others might consider a main road. It doesnât make sense to have such a narrow definition, use what tag feels ârightâ. highway=path particularly has a very broad meaning, other than meaning a way to get from A to B not much else more can be implied.
I commonly encounter something that was originally a highway=track but over time it has become overgrown due to not being used by vehicles.
If it is just slight overgrown but thereâs still the two wheel tracks visible and a vehicle could still push through over the light vegetation Iâll leave it as highway=track, but if itâs severely overgrown and thereâs only one worn path left from walkers or cyclists, Iâll convert it back to a highway=path or highway=footway to reflect the current on the ground state. Or more correctly you could include abandoned:highway=track alongside the new highway=* value to indicate it used to be a track but itâs been recently abandoned.
If later on it gets clear and restored back to a usable track then you can update it back to highway=track. Having the tags changing like should be seen as a good thing, it this helps to keep the data current and accurately reflecting the ground truth.
Nothing is ever static, being declared, built and used as a track doesnât mean it will forever hold that status if not maintained.
The wiki already avoids picking one: highway=track just needs to be wide enough for a general-purpose 4WD to physically fit, and that width isnât fixed, it depends on what vehicles are actually used in the region (see this footnote on highway=track). A single number in the global docs would be wrong.
In Thailand that question is often literally unanswerable at a single point in time. A trackâs condition changes constantly, between seasons and over years: some become overgrown or impassable during rainy season, unrecognizable as a track, then get bladed clear again right after. Others get abandoned for a few years, or for good. Whichever moment you happen to survey it, âtypically usedâ can go either way depending on the season, which is exactly why width works better as the anchor, it doesnât change with the weather.
Right, and itâs not really about whatâs âtypicalâ to see driving there, itâs the minimum a vehicle needs to physically fit through. We settled that for Thailand a few years back (topic 4648): 2 m, the minimum width for a pickup. Still documented today.
Thatâs the case for keeping obstacle=vegetation on a known highway=track, an obstacle that eventually clears, however long it takes, isnât really a reclassification, just an update once you see itâs clear again.
Thatâs roughly the split weâre circling: an obstacle that eventually clears stays highway=track with the obstacle tagged, one that never does becomes abandoned:highway=track + highway=path. Where weâre stuck is telling those apart. A duration threshold sounds right in theory, but a mapper whoâs only seen the place once has no way to confirm it.
One open question for the thread on that: would a visual double-track signature, worn ruts, an overgrown center band between two still-clear wheel lines, or twin concrete strips with grass down the middle, count as decent tie-breaker evidence in the disputable/overgrown cases?
Every highway=track is wide enough for 4-wheel vehicles, so itâs a necessary condition. On the other hand, not all ways that are wide enough are highway=track (they could be e.g. highway=primary).
A sufficient condition could be âlow traffic, primarily agricultural/forestryâ. If this is the case, then itâs definitely a highway=track (suitable width included), but if not, then it might still be, we donât know.
For me, the criterion for highway=track is some combination of low traffic and created for four-wheeled vehicles (with other traffic as an afterthought).
Iâm not sure if Iâm interpreting this right, but I feel like you may be judging against an implied standard of âhow would we tag this if we had perfect knowledge of all relevant factors in space and time?â
It would be nice if all rural ways could be surveyed by an army of OSM mappers familiar with their usage and physical condition over all seasons and several years. But we do the best we can with the resources available. And yes, that often means we have to make a judgement based on a single visit. Often a visit during which we ourselves are the only visible âtrafficâ.
The alternative, for many tracks I survey, would be to leave them forever at version 1 which was drawn, or guessed, from Yahoo imagery in 2009 (aside from tagging, the geometry often needs to be redrawn). So I tag them as tracks if they look like tracks to me, and paths if they look like paths. Maybe for 1 in 100 edge cases Iâd make a different decision if I visited 6 months later. Itâs still a lot better than not surveying at all.
.
The post of mine may have been a gut reaction to the advise given by lots of people here when presented a photo of a track with lots of non-woody vegetation and a small trampled path in the centre. For me it is obvious, that a tractor can drive there with ease and for a certain will do so, when the next logging campaign due. For some it is obvious that this is too narrow for a vehicle.
The subject is packed with implications: Tracks here mostly private, Iâd say, exclusively so, mowing the vegetation no problem for the land-owner, turning a path into a track requiring administrative pardon â will have to bulldoze and grade and so on. I am a local, I know that stuff. Sorry for being obnoxious. I do not think this goes as far as having perfect knowledge 
Common knowledge perhaps?
PS: The sample I mentioned above, I then just slapped tracktype=grade5 on it. I was not aware of the obstacle=vegetation and other tags. I still think, the routing signal should come out the same? BTW: Only pedestrian public allowed there, but this fact not reflected in the OSM data, but a StreetComplete user absurdly added the completely unexpected fact, that lit=no there on this grade5 track in the backwoods in the mean time.
Yes, and thatâs one reason adding a specific width to the global wiki may not be helpful. For example, it seems that you are implicitly considering a kind of âavailable width if easily cleared vegetation is removedâ. A decision based on width will often be easier for a way that passes between two immovable stone walls than one crossing grassy terrain.
You also mentioned ownership of the track having an influence. It has been mentioned on this forum before that mappers in some regions may have a concept of a track as something very likely to lie within private property, while in other regions tracks running between private property boundaries are common. I mostly map public tracks, both because they are common here and because private tracks here are often really private, i.e. closed to pedestrians as well as vehicles.
In the end this seems to come down to a slight ambiguity between paths and (typically) grade5 tracks in regions where tracks may periodically become overgrown in between uses by vehicles. Iâm not sure if the decision one way or the other has any significant impact on map users.
A disused track which has a clearly visible path running through it could justifiably be mapped as either.
I like the idea of having both, with the âtrackâ aspect hidden behind a lifecycle prefix. This carries implications that there may be uncleared fallen trees and other deterioration that might not have been present when the wayâs tagging was last modified.
Worth separating two things: surface=grass just describes ground cover, a track fully covered in short grass is still a track at full width. obstacle=vegetation is the case that actually narrows what fits, tall grass, branches, bamboo, and thatâs where the current passable gap, not the as-built width, should decide highway=track vs highway=path. Clear it and itâs a track again, same as your stone wall except temporary.