I agree. At the same time, I think maps are oriented towards a specific audience, and usually geographically limited. In OSM, I think we try to find schemes that work around the globe, while still enabling the rendering of the specialised maps for different audiences and regions.
The legends could show the differentiations the maps want to make, but we don’t have to tag these particular differentiations as categories or values. We can expect the renderer to select and combine tags, reducing the data to what they want to show. It would be good to make that as easy as possible, though.
So we can tag a dozen surfaces while the renderer only wants to show paved vs unpaved, but we do it in one key with values covering the world, with clear values which the local renderer can interpret as paved/unpaved according to what this particular target audience in this particular region wants, while the global renderer can use a different surface rendering legend.
If we’re comparing map legends, then here is Ordnance Survey, Scotland, 1:25,000. No attempt to show the difficulty of paths. (According to the publisher, these maps are “ideal for activities like hillwalking, hiking, mountaineering, trail running and mountain biking”).
Here’s the original thesis where that legend appears. Apparently it’s from an IGN TOP 25 map. Do you know if this is typical of French topographic maps?
What I find interesting about this legend is that it’s an entire scale of… not difficulty per se, but usability more holistically. To replicate this legend using OSM data, one would need to employ not only one of the keys like trail_visibility=* but multiple in concert. Or maybe “hard part of hiking trail” is relative to the trail itself rather than a universal scale?
Agreed. In other threads, I often come out against attempts to design bespoke typologies or abuse third-party scales, mainly because these attempts quickly turn out to be parochial and inflexible. (Hello, crossing=* and isced:level=*.) On the other hand, I do see value in intuitive, keyword-based keys to supplement more empirical keys about specific physical characteristics. The point of studying map legends is to ground the discussion: is a distinction so fundamental that it affects the very language that people use refer to the thing, even if they know nothing else about the thing?
One thing I’ve learned in studying legends for OSM Americana is the degree to which print cartographers play fast and loose with any physical characteristics they do mention. A seemingly straightforward scale of place labels sized by population – yet some cities actually get boosted or demoted based on notability. A seemingly precise specification of inclines by grade – but a reader will never fact-check the map for slight deviations from this criterion.
Quite interesting, considering the title says “roads and paths”.
If we were to only take intersection of different legends, we’d end up with an empty map. With union, there would be a detailed map.
Here we have trail and primitive trail distinguished by two different dash patterns, plus a special yellow highlight for the Pacific Northwest Trail. A parenthetical tells us that a primitive trail may be difficult to follow, so we can assume that the other trail classification should generally be easier to follow. Here is a sample of all three line styles on the map:
Now here is another area of the map where the Pacific Northwest Trail follows the South Coast (below Third Beach) and the North Coast (above Rialto Beach). Some sections of this yellow highlighted line have long black dashes indicating trail or short black dashes indicating primitive trail, but other sections lack a black line of any kind. These sections have no visible path and the hiking route simply follows the shoreline.
This is from the IGN 1:25,000 that is still used as THE reference here. But this belongs to the “touristic overlay” that does not belong to the topographic map itself (it is compiled from sources such as mountaineering and hiking associations)
I think every single topographic map of the Alps I have ever seen (Austria, France, Slovenia - I have not used paper maps in years and do not have them on hand), and also both French and Spanish maps of the Pyrenees, had full, dashed and dotted lines to distinguish easy paths, intermediate paths and hard (hand-using) paths. To me this is the gold standard :-).
This looks a misunderstanding to me, Austrian maps by ÖAV and BEV use stipples for width, so much it seems to me. I posted a photo of a legend of a printed map above. You may read it as something different, but the legend is what counts, for me.
This is the Austrian and German classification scheme for hiking paths Blue, red, black. Obviously, not good enough for openstreetmap though.
It absolutely makes sense for producing a map to fall down to a lower number of categories that we are able to map. For path, I would be happy with 3 already.
That would admittedly be awesome for several reasons. If we manage to do that in the summer, I could do that.
That region though, is in Trentino, not South-Tyrol, for what it’s worth.
All the (2) people familiar with the surroundings agreed that there is nothing on the ground that guides the way of someone passing there, no markings, no tramplings, just ragged terrain. Mapper CAI_Salò in addition said, that the Kompass map as well as the OSM mapping give indications on where to head, for use “by proven expert mountaineers, by azimuth”.
I do not think having commented “When there is nothing on the ground but there is a route in your head helped by a line on a (paper/on-screen)-map, openstreetmap data should also only contain a route and not a path” would have met a different reply.
My take home from this: For some a route (an idea) is a kind of path (type=route perhaps.) From various replies in several topics, I started to learn, that in English (both British and American) that sounds quite reasonable.
