I have read, openstreetmap data is growing incrementally. First, people just map the bare geometry. Then others chime in and add the details. Recently I observed this, Tödliche Gefahr auf falschen Wanderwegen - Zillertaler Zeitung – the website shows three pictures of what roughly got mapped as a bare highway=path with no attributes.
I am a bit perplexed: Why do consumers of openstreetmap data treat such an unattributed path as suitable for “a walk in the park”, i.e. something any tourist can manage without special equipment or skills or you name it prerequisites?
I am even more perplexed, I post a link to press where people died because of trusting openstreetmap data that is obviously wrong at least outside of “openstreetmap guiding principles“ and the post gets a popcorn emoji?
I reacted with popcorn because the topic of tagging hiking paths has been discussed repeatedly and recently, including by you referencing hiking deaths in Number of fatalities and sac_scale in July, and I haven’t seen anything that would indicate the discussion here would go any different
Additionally your main question is rhetorical and would be better addressed to developers of the apps you have in mind rather than here onto the mappers forum
I couldn’t help thinking “Here we go again”, and almost reacted with a popcorn emoji too. Previus discussions were endless. Solutions have been discussed, but no-one has made a proposal.
because they are not aware of the context (an unattributed way in the park looks the same to them as such a way in the wilderness), and because we expect difficult ways to have additional attributes, e.g. sac scale or trail visibility etc.
Well, the popcorn emoji is notoriously ambiguous. My take is that this isn’t a problem specific to OSM, or really to paths either. The first link in a simple web search returned a case where a person was killed because they drove their car over a collapsed bridge trusting Google maps.
At least in OSM, people who notice incorrect or imperfect data on the map can correct the data themselves. Or leave an anonymous note, with a much better chance of getting the issue fixed than with Google maps (the story I linked to claims that Google was notified “years before”).
Regarding paths specifically, the posts above me nailed it, @dieterdreist in particular. As @Peter_Elderson noted, solutions have been discussed, and the proposal to treat bare paths as vague and imperfectly mapped until people add more tags to them seems to have won for now. Rome wasn’t built in a day. In the interim: how about treating barepaths outside residential areas or parks (or x kilometers from a place=city node) differently, i.e.: as potentially dangerous or impassable, in the routing graph?
In that case, almost all paths in Nederland would become potentially dangerous or impassable, where in reality you can walk them blindfolded drinking Heineken and without spilling a single corn flake.
Article seems to not be mentioning OpenStreetMap, so I am not really sure is it OSM-specific (though I relied on autotranslate)
You just linked article in German, without giving context
People likely seen another highway=path thread without anything new, with rhetorical question that you almost certainly know answer to
Based on image, what trusting OSM data would get them is an irritating detour and a spoiled trip (assuming that it even was OSM data and photo is actually related). Walking on highly risky and obviously risky and clearly risky path is not really OSM fault. If someone would drive car into a flood because GPS told them to drive into a freshly fully flooded tunnel it would not be entirely or even mostly OSM fault either.
you have not even identified relevant OSM object
And for question stated in the OP? refer to previous path threads
I do not get reason for starting a new path thread that is not even proposing something theoretically actionable.
Because your claim is conjecture, maybe they were using OSM data, maybe they were not, maybe they were “trusting” OSM, maybe they just used it as a hint that there was some kind of path there. We simply don’t know at this point.
The other aspect is, because hiking is (in absolute numbers of fatalities) the most dangerous active pastime, there is a constant supply of accidents to speculate about, trying to turn every one in to an OSM scandal is just not particularly helpful.
Sorry for not making this explicit, but the “x” in my post above was meant to represent a variable. And the whole proposal was meant to be an example of a simple heureistic trying to compensate for imperfect data.
How many paths are there in Netherlands that would be, say, 30 kilometers away from a city-node? Of course geography plays a role here. The area of Netherlands is almost ten times smaller than Finland with roughly thrice the population. So far away paths are probably more prevalent in countries like Finland (or mountainous places like Switzerland) than Netherlands.
I don’t think that that’s really true - lots of people have been banging on for years about possible ways forward (here, here, here, here, here etc.).
My own take is that “highway=path without other tags is unhelpful” and it’s always best to try and capture other relevant tags (which may include any of surface, smoothness, sac_scale, trail_visibility etc. etc.).
it just won’t work at all. First it has nothing to do with “city” (i.e. big towns), because you can find harmless paths in smaller towns, villages and hamlets just as well. And in the mountains, it is quite typical to have “dangerous” / “demanding” paths also close to settlements. Or even inside the limits of big cities, there can be dangerous paths, e.g. in former quarries.
That is true, and I did not mean to discard such efforts. At the same time, I don’t recall any formal RFC/CFV procedures, e.g. to formalize a path=… subdivision, containing e.g. path=mountain_track.
If you mean “some waffle on the OSM wiki” then (a) you may be correct (although sure people must have tried) and (b) I suspect that it’d be a complete waste of everyone’s time.
The tags that people use are determined by the information that they have about the things that they are mapping, and how the editor that they are using guides their contributions. Editors tend to reflect usage, what other people think in issues and discussions, as well as “what the last person to update the OSM wiki thought”. They don’t just reflect the last of these (thankfully).
The discussion around the highway=scramble proposal must have been so traumatic that everybody involved seems to have forgotten about it. Does that mean I can just recycle it and see if three years later it might pass? It still is a good idea.
