I’m very confused with tracktype and other similar tags, and I am unable to find a clear explanation.
The OSM wiki states that “tracktype is a measure of how well maintained a track or other road is”. However, the comments in the values table seem to suggest that tracktype is a measure of firmness (ranging form hard/paved/sealed to soft/uncompacted surfaces) which is confirmed by the smoothness page that states that tracktype “refers to surface firmness, regardless of shape”.
So, let’s assume that well maintained tracks are generally firmer than unmaintained ones to connect both concepts. In this case, there are some marginal but not uncommon real-world cases that I find difficult to graduate:
Bedrock tracks. They are intrinsically unmaintained but generally very compact. They can be sometimes very smooth.
Hard-soil tracks with loose less-than-fist-size stones that do not compromise vehicle traction.
Maintained deep fine gravel tracks (loose by nature) which are fine for cars but problematic for bicycles.
You’re touching upon some of the inherent weaknesses of “quality” tags like tracktype and smoothness. There is an endless variety of road and path surfaces that vary in quality and also depend on the vehicle you use.
The confusion about tracktype may be caused by its history: it started out as a measure of passability of tracks (now covered by smoothness), but evolved towards a measure of surface composition (mix of hard and soft materials) and then as a measure of surface firmness. I would continue to use it in its latest meaning, i.e. surface firmness.
Definitely tracktype=grade1 (even though it’s not paved)
tracktype=grade2, because you state it’s “hard-soil”
That’s a tricky one: the last word on how this should be tagged hasn’t been written yet. I’m a car-centric person so would tag it surface=fine_gravel + smoothness=intermediate (probably) + tracktype=grade3 (even though it’s not a mix of hard and soft materials, fine gravel by itself is medium-soft for broad tires). But a bicycle-centric mapper may tag it as tracktype=grade5 because it’s as soft as mud for his bike. That it’s hard to ride on is a property of the surface itself, so maybe we should invent a new surface value for it: surface=loose_gravel (and surface=loose_sand, with implied meaning “hard to ride, but no problem to drive”).
So your recommendation is to use tracktype as “perceived” firmness, ignoring maintenance status and track composition mixture (hard/soft) as suggested by the wiki, isn’t it? It seems logical to me.
In my home area most of the tracks are naturally compacted (a hard mixture of sand and rocks) but partly covered by loose fist-size rocks. By car, these tracks feel harsh and firm. Thin bicycle wheels don’t sink. I usually tagged them tracktype=3, mainly because fist-size rocks didn’t seem to fit the “gravel road” definition for 2, and there were hard enough to rule out 4 or 5. Then I added the smoothness and mtb:scale/:uphill tags to let car and bicycle users figure out how passable the track was. Should I have tagged them tracktype=2?
But a bicycle-centric mapper may tag it as tracktype=grade5because it’s as soft as mud for his bike.
Hence, are we discussing floatability when we talk about softness?
The idea that tracktype is about firmness really only makes sense in places where the ground is naturally soft. The tag has always been a rough classification based on a few different factors but essentially boiling down to how well developed the track is. On the top end a grade1 is a well developed road and on the low end a grade5 is a faint imprint on the landscape that barely qualifies as a road. It does not make sense for a faint track to be tagged grade1 just because it traverses naturally firm ground.
My understanding of tracktype is very much the same as explained by @ezekielf. The different values reflect the stage of development work invested and (less important) maintenance.
Grade 5 nearly undeveloped: the only work done was to remove vegetation, flatten some bad bumps and holes and remove big loose stones.
Grade 4 development includes some levelling of the complete track, removing more-than-fist-size-stones, adding some gravel here and there to stabilize soft sections, but not much more. Heavy machinery may be used but to a limited scope only.
Grade 3 development includes full levelling and a certain degree of compaction of the track to create a more or less consistent surface. Heavy machinery has to be used normally.
Grade 2 development includes a solid substructure (mixed gravel of different granulation) and a fully compacted surface. Construction not possible without intensive use of heavy machinery.
Grade 1 development finally adds a pavement as top layer.
More details are added by surface=* and smoothness=* as you are aware already.
hard soil tracks tracktype=grade5/grade4/grade3 + surface=ground + smoothness=*
deep fine gravel tracks tracktype=grade3/grade2 + surface=fine_gravel + smoothness=*
The last smoothness depending on how “deep” the fine gravel layer is. According to my own experience deep fine gravel tracks don’t remain “deep” for long time, specially when motor vehicles are driving there, as fine gravel tends to be brushed aside by fast moving vehicles quickly.
I strongly disagree with @ezekielf and @Map_HeRo : it is in disagreement with the wiki and would be a complete redefinition of tracktype.
I think that how “good” (well-developed, maintained,…) a road is should be expressed by smoothness. A well developed/maintained, i.e. newly built road is likely to be with a high degree of smoothness (paved with asphalt or concrete is likely to be smoothness=excellent, surface=paving_stones is probably less smooth when new, and surface=sett is likely to be less smooth than that. Due to use, a road deteriorates, which can be expressed by tagging it with a worse smoothness value, until maintenance comes along and the smoothness is improved.
For unpaved roads (tracks), it all depends on the amount of effort that has been put into building the road. Has additional material such as gravel been added, has it been compacted or just bulldozed, or was there no building at all (a track formed by driving on a natural surface)? Again smoothness is a measure of the quality of a surface, and additionally it can be tagged with tracktype to express the firmness of the surface (which is related to the mix of hard (stones) and soft (sand, clay) material in the surface, with more stones leading to a firmer surface). I think tracktype has limited practical value (I usually don’t bother to tag it), but could be used as an estimation of how fast the surface quality changes upon use, wet weather, etc. as well as the risk of getting stuck on a track because your vehicle sinks into the surface. tracktype=grade1 is very solid, so very unlikely to change due to use or wet weather, while tracktype=grade5is very soft, so after rainy weather it will likely be very muddy, and a heavy vehicle is likely to get stuck.
A look back at history: Proposal:Grade1-5 - OpenStreetMap Wiki before it went to vote (which never happened.) Will openstreetmap ever get as simple as that? The pictures quite telling, at least for what I observe in the area of my local knowledge.
If tracktype only maps firmness, then perhaps it is mostly meaningless in your area. Can you share a photo or two?
The hedges prominent in description, less so in the photos, actually only grade 3 and 4 segregated by hedges. Perhaps a very locally grounded concept, still applies quite well outside of Britain too, e.g. here where I hail from, likely not all over the world, though.
The photos still feature in the main documentation. While they are a bit about firmness, I still wonder why tracktypeshould map firmness only, shouldn’t the key then be renamed info surface_firmness? The pictures definitely also show the different efforts that went into construction; an objective measure, not?
I was under the impression that smoothness is an indication of the minimum robustness of wheels needed to use the way, e.g. city racing bikes with thin tires cannot accept anything lower than the highest value, normal bikes can use the 2nd highest value, cars can use the top 5 values, etc.
Also, smoothness= should be tagged as the worst part of the way, while tracktype might be more of an overall average?
Yes, that’s how road surface quality is expressed. An excellent road can be used by any wheeled vehicle, while a road that can’t be used by most vehicles is very bad.