Ideally path would be abolished worldwide :-). It is hard to judge from one picture - if it is this wide and smooth all the way, I would seriously consider track and auto=no (track again is about what it looks like and how it it can be used, not so much about legality). It does not look built at all so I lean to trail, but again, hard to judge from a picture. is it on streetview or something?
1 Like
If you talking about highway=unclassified..primaryā¦ Someone invented them in times where āsecondaryā tags were not existing aiming to show different level of importance. In the end those differences are kind of useless. As they are globally not aligning. A primary can be a 6-lanes asphalt road or can be on the level of highway=track. Modern road-maps evaluating the ref or the route-relataion and their network-tag. 
3 Likes
[citation needed]
More seriously, what maps show (even road maps) will depend on their target market and what they want to tell people about.
Want to show relative position in a hierarchy? Then primary, secondary etc is likely what you want. Want to show condition? surface, smoothness etc. are better. Want to show route relations in the situation where your country has multiple routes over the same bit of tarmac? Show those. If your country doesnāt have those**? Donāt.
** mine doesnāt
1 Like
Take a look at OSMAmericana
For sure what a map is evaluating depends on their purpose. Iām also not saying, we have to remove those. I was just asked, why they exist. They working well in areas where you have 4 defined levels of importance. In areas where thatās not the case, they are as well rather random used.
So when we do the ad absurdum argument that we could just as well have only highway=road and everything else would be in secondary tags, that would actually be a preferred state of things for you, ha 
?
Iām not saying that. All Iām saying is, that it could have worked out as well. Everything could have been highway=road and there could have been another tag level_of_importance. Same as it was done for railway where you have railway=rail and usage=* or service=* for whatās the purpose of that rail.
2 Likes
I sympathize with splitting path/trail. Iāve seen via_ferrata mentioned in a sibling topic. Mind you: There is no progression in this list, especially not from the easy to the hard. Trails (as the term coined above) can be much more challenging than via ferratas.
Apart from that, the so-called true hw=path (<5% of paths here) certainly not even worth sac_scale=strolling.
1 Like
First of all, after reading through the posts up to here, I would suggest to talk about generic ways (as in generic, non-motorized ways) and not about shared_use (which might be misunderstood as MUST be shared).
The right one is hw=generic with two access tags.
The left one is hw=generic with no access tags, because it is wide, well-built and maintained and I donāt see tractor tracks whicht would make it a track. This is exactly the type of way with cannot be tagged properly/distinguished from an actual trail today because both are currently hw=path without access tags
1 Like
From the statistics, I would derive that the 1% of path-with-access-tags should be replaced, so we get something like
hw=generic - constructed/maintained ways for non-motorized traffic
hw=path - real single trail hiking paths, further described with [ā¦]_scale
hw=via_ferrata
But realistically I do not believe that any use of path can be changed. 1% ist still a lot and it impacts all editors and maps.
I think the only realistic and backwards compatible way out of this dilemma ist to require sac_scale for real hiking paths and prohibit it for mixed foot/cycleways. It is already well established, just sac_scale=hiking is not tagged but rather implied. And was never intended for plain compacted/paved ways.
So we would get:
hw=path : generic use
hw=path + sac_scale : actual paths
hw=via_ferrata
Thatās a misunderstanding: as @Hungerburg has been pointing out repeatedly, a via ferrata is a really specific thing, so we should not use it as a generic tag for climbing routes. The generic tag for climbing routes could be climbing=route but itās not really clear how it should be used.
1 Like
Such distinctions wonāt work in most of the developing world for a few reasons:
-
Physical distinction is irrelevant: You can find well-constructed hiking trail segments, while urban multi-use paths might look like rugged trails.
-
Non-motorized distinction isnāt practical: Multi-use paths are often legally accessible to small motorized vehicles, like motorcycles, by default. At the same time, trailsāusually thought of as non-motorizedāare frequently used by off-road motorcycles and ATVs, even in Western countries.
The distinction should be based on function and usage, rather than legal access or physical attributes. For instance, a path used primarily for commuting should be categorized differently from one used mainly for recreation, regardless of whether itās accessible to motorized or non-motorized traffic.
1 Like
I strongly disagree: the main usage of a way is highly subjective and not verifiable because we donāt have access to information on why a user of a way is using it (weād have to interview a significant number of users and ask if they are on their way to work or just going for a walk, etc.).
I understand that distinctions between paths have always been made based on legal access, and I think we should continue to do so. We should look at regional and national traffic laws and local signage: if a path is designated (by law or signage) for pedestrians, itās a highway=footway, if itās for bicycles only, itās a highway=cycleway, if itās designated for both, itās a highway=footcycleway (or something similar), etc. and only if there are no rules, itās a highway=path.
The highway=path category is then a large one, but I donāt think we can divide it further based on objective criteria. The best we can do is to describe the various highway=path with secondary tags describing their quality.
1 Like
The commuting vs. recreation distinction was just a suggestion, we already deal with subjective, non-verifiable aspects in other road classifications (like which one has more traffic or is considered more important).
I think the legal distinction is one of the main sources of conflict, and I donāt get why pathways were given special treatment in the first place. Thereās no reason it couldnāt be changed, especially if a new pathway=* top-level tag is introduced at some point.
That would mean weād need a ton of new tags for designated small vehicles, like highway=mtbway, highway=motorcycleway, highway=atvway, and highway=golfcart_way. But proposals like these are unlikely to be accepted.
No, it doesnāt and we donāt. We already have foot=yes/no, bicycle=yes/no, segregated=yes/no and bicycle/foot=designated as additional tags to highway=footway and highway=cycleway, so there is no need e.g. for a footcycleway top-level tag.
This very same principle appliesāthough mutatis mutandis of courseāto highway=paths. We have descriptive tags like width=, surface=, smoothness=, trail_visibility=, informal=yes/no, etc., and even motorcar=no for those paths that need it.
Notice that none of the additional tags above are strictly speaking subjective (well, there are endless arguments between smoothess=bad and smoothess=horrible, but either one provides a rough idea (pun intended) on the surface smoothness), but describe objectively verifiable properties (local laws, traffic signs, surface markings and materials, widths, etc.), pecisely as @rhhs mentioned above.
Sure, but ideally the tagging should express the peculiarities in OSM-wide terms, so all know what the situation is.
This is not the citation you are looking for. Like any other OSM based map, Americana uses the highway=* classifications values to choose an appropriate minimum zoom level for each road segment. The classification is also one of several factors affecting the line color and weight. Route relations with network & ref are excellent for displaying highway shields. They are not a replacement for highway=* classification values.
2 Likes
Wouldnāt it be fair then to deprecate and replace highway=footway/bridleway/cycleway with highway=path + foot/horse/bicycle=designated?
Getting rid of this double standard would definitely help.
3 Likes
Sure we do - from survey.
People should get out more 
5 Likes
While the details depend on local legislation (e.g. Freedom to Roam laws, etc.) itās far from obvious that it would. Help, that is.
The top-level tag (i.e. footway, cycleway, and path) is able to immediately express the existence (and in the case of a path, the absence) of a roundel-type traffic sign that entails legal rights-of-ways. Further sub-types of roundels can be expressed w/ segregated, foot/bicycle=no/yes etc. designated expresses the existence of non-roundel type traffic signs (that do not engender rights-of-ways), even on paths. In places where FtR laws do not apply, foot/bicycle access-tags are essential especially for paths, in the FtR wold they are implied for them.
The combinatorics here works to our advantage in that we are able to express a very wide range of possible legal (and/or practical) categories while not having to have to have dozens of top-level tags for this while having a reasonable set of top-level tags to express the (objective) main purpose/type.
While I do not argue that the tripartite top-level tags (cycleway/footway/path) are perfect and forever immutable, I would say that they are a reasonable compromise and absolutely do reflect a fundamental, basic ground truth in much of the world, for most of the ways that are primarily intended non-motorized travel.
1 Like
How close to being an achievable goal do you think that is? Based on all the discussions around path that there have been (since around 2007) how likely is that ever to happen?
If OSM was starting from scratch then Iām sure that there would be lots of changes made as to how some things work (Iām sure the wiki wouldnāt exist in the same form, for example). However, itās not the mid-2000s any more. The contributors to OSM arenāt a few people around the table in a pub in London; theyāre significantly more, spread all around the world. You (and I) donāt get to tell them how to tag things; we can only make suggestions using reasonable arguments.
Any plan to change the way we tag things as a community needs to include a plan to explain the change and persuade people to adopt the changeā¦ It also needs to plan to change the data, but frankly compared to persuading the OSM community thatās the easy bit.
3 Likes