Effect of oneway on pedestrians?

Help and support Tagging help & support
, ,
101

Ok, maybe it does not even matter. As long as the wiki explains sufficiently well that ambiguity needs to be dealt with by applying an extra tag and the mapping community follows this suggestion (which isn’t trivial at all) people could even use either of the two tags. Two different tags with the same meaning won’t be an issue for data consumers at all. And if after a while one is used more frequently than the other one could still be deprecated? I’m not saying this is ideal, but two tags are better than none, if you ask me, and maybe having a choice even helps with the adoption.

I know :slight_smile:

102


|
|

  • | - |

So you would go all the way with the “oneway only applies to vehicles” rule and also discourage using it for aerialway, steps and via_ferrata for example? That also makes sense to me, but I’m not sure it would be possible to deploy such a change? So I wonder if as a first goal we should focus on improving the situation for the trickiest cases where it is actually hard to tell if a oneway was meant to affect pedestrians or not (especially highway=footway, but also path and pedestrian)

I would say that for “highway=pedestrian” it is unquestionable that it applies only to “vehicles”, as vehicular exceptions are expected and oneway restrictions are common (at least around here). Applying these to pedestrians would just be utterly wrong and cause lots of trouble.

Something that maybe wasn’t said yet: according to the Vienna Convention on road traffic, oneway is a restriction applying only to vehicles. I know that not every country has ratified it, but many have.

For steps I am not sure, but I have definitely already seen steps that were drivable (and people were driving on them), so these may be similar to highway=footway. I have queried overpass api for oneway=* on highway=steps and the result were 6804 ways in the world, out of 1728000 instances of highway=steps, refining the search to look only for oneway=yes, it still were 3836 ways and 36 ways with oneway=-1, (2921 of the oneway tags on steps are just stating the default oneway=no), i.e. 0,2% of steps have a positive oneway tag and would have to be looked at. Looking at some examples, e.g. https://www.openstreetmap.org/way/134605400, these were particular situations where there isn’t actually a oneway restriction on the ground, you could legally walk back and forth up to the barrier, but you can only exit and not enter the place (in this particular instance). I guess many of these are similar situations were the oneway is questionable and not actually a legal restriction, but a physical one (turn stiles or similar).

For aerialway=zip_line I wonder if there is any chance they cannot be unidirectional? The wiki says it is a “gravity propelled aerial ropeslide like a flying fox or similar”, how could gravity make you travel from the bottom to the top?

Regarding oneway:foot, I think the vehicle:foot analogy makes it clear. It just isn’t logical to subtype a tag that doesn’t apply anyway.

I’m not sure I understand the analogy. Sure, overwriting vehicle=no for foot goes like foot=no, not vehicle:foot=no. Or do you mean this does not ‘overwrite’ anything because foot is not covered by vehicle in the first place?

yes exactly, as vehicle does not apply to foot, vehicle:foot=anything does not make sense.

103

I should better refrain from discussing the specifics of the English language with a (supposed) native speaker, as I am not, but from looking in the Internet, it seems as if driving a carriage is a term that could be used somewhere where English is spoken.

I guess the word “drive” has been removed in favor of stating it applies to “vehicles” exactly because of this, because you do not “drive” a bicycle, but a bike is a vehicle and the restriction applies to it.

1 Like
104

I would just retag these so we don’t end up with 2 different tags for a directional restriction on highways based on the highway value. “foot:backward=no” is already defined (generally, nothing special about it, it follows the generally established semantics), it does not suffer from the problems oneway does, it is unambiguous, why the resistance?

105

You’ve just posted three separate consecutive replies to three different people. Perhaps the fact that you’ve had to do that suggests that your view is rarer than you think it is and is not, in fact, the consensus? :grinning:

1 Like
106

oneway restrictions are also appearing on objects not covered by Vienna Convention on road traffic

especially oneway restrictions applicable to pedestrians

in particular, mountainous trails with oneway pedestrian traffic are not covered by it, as it is not road traffic

because it mismatches already established and used oneway:* schema, is basically impossible to remember and mismatches also actually used oneway on purely pedestrian ones

1 Like
107

While those on first sight may be obviously oneway for pedestrians, they both suffer from problems of

  1. side effects: if a mapper comes later and changes via_ferrata to path for the former, or removes bicycle=no from the latter, then suddenly oneway=yes changes meaning.
  2. complexity: data consumers have to look at bunch of other tags to decide which mode of transport oneway=yes pertains to. Tags should be as unambiguous as possible.
2 Likes
108

This! We should not allow a tag like oneway=* to mean different things depending on whether you put it on a footway, a cycleway, or a road. which makes the least problematic solution that it applies to vehicles only, and a different tagging is needed for pedestrians, nordic skiing, and possibly horses.

But I think the :<direction>=no solution is probably not the most intuitive one, same goes for oneway:foot=yes. So maybe something completely different is called for, which applies to all traffic (something like unidirectional=yes), and not some awkward way to make it work for one mode of transportation additionally.

1 Like
109

To avoid ambiguity highways that pedestrians can only use in one direction should be tagged with a dedicated tag other than a simple oneway=yes (how this tag reads is still up to discussion and not specified in this poll).

Explanation: The ambiguity arises, because in the majority of cases like highway=residential,tertiary,service etc. oneway=yes does not affect pedestrians. At the same time it is hard to identify highways that are only used by pedestrians (in which case the ambiguity might be resolved) consistently now and in the future (upon further editing).

  • yes
  • no
0 voters
1 Like
110

On this part of it, I’d suggest that the following are likely to be only used by pedestrians:

  • highway=footway, unless there is some other tag (perhaps horse=permisssive?) that suggests that other transport modes are supported.
  • highway=path with a foot access tag but without some other tag (perhaps mtb:scale=4?) to suggest that modes other than foot are also supported.
  • highway=steps with a foot access tag but without other tags to suggest that modes other than foot are also supported.

To be clear, in each of those cases and in the absense of other information I’m suggesting that:

  • Probably only foot as a mode is supported
  • A oneway=yes in each of those cases would logically apply to foot.

There will be exceptions to all of these, of course. A country such as Scotland** allows both bicycle and foot access in most cases to things we’d map in OSM as a highway=path. A quick search finds these. For each of those overpass query results, I’d suggest that more clarification about which modes of transport are actually supported and/or which modes of transport the oneway applies to would be good.

** legal access details

1 Like
111

Would you find it distracting if there was an additional tag that indicates the one-directional attribute for pedestrians explicitly in these cases, even though it might be possible to recognize this based on other tags?

112

Given that oneways for pedestrians are relatively rare, requiring an additional oneway:foot=yes to avoid the ambiguity seems alright

1 Like
113

I state this again, because I don’t think this is the solution we are looking for: if oneway=yes applies to vehicles only, you will need exceptions for horses (which are not vehicles and highway=bridleway + oneway:horse=yes feels awkward), and skiing (yes, we have lots of nordic skiing routes). While I think it makes sense to have an exception for skiing (because generally, you can only go downhill, and loipes are often found on top of paths which are not oneway during summer), it would be better to have a “this is one way for every mode of transportation”, because usually, if a way must be used in a specific direction by pedestrians, this applies to all other modes of transport as well. At least that’s my observation in the parts of the world I’ve been so far. I have yet to find a way that’s oneway for pedestrians, and non-oneway for others.

114

This sign makes it not oneway. The behaviour of people may be as expected, intended by setting this traffic_sign.

Because like similair signs with vehicles on it.
When you pass this sign, you may not use the carriageway/path surface behind the sign, a rule only from this side entering.
But from the other side there is not such sign, then when they walk(drive/ride) in, they can turn around and walk(drive/ride) back, legal, so setting foot:forward=no on the whole path is not correct. In the middle of the path there could be a tourism point. To navigate to and navigate back. 10 meters of foot:forward=no after this sign is more correct.

1 Like
115

Yes, it’s a classic example of wrong signage, because either, they didn’t know better, or, like in this case, there is no official sign to make a path oneway for pedestrians. We have mis-uses like these over here as well, when they want to forbid driving on a cycleway in the opposite direction :man_facepalming:

To my defense, I didn’t add this example to the wiki, but I also think that the intention how the way should be used is understandable.

1 Like
116

Personally, no I wouldn’t, but (a) it will be a challenge to get every other mapper to do that and (b) where there is genuine ambiguity it will need an on-site survey to set the other tags needed to remove the ambiguity.

117

How is highway=bridleway+oneway:horse=yes more awkward than highway=footway+oneway:foot=yes?

If this is the case there is no issue, right? If there was a oneway:foot=yes tag you would automatically know that riding in the other direction is forbidden, too. It might be a bit more complicated if there is foot=no already. But then you could just use oneway:horse=yes (or to take this even further oneway:access=yes, edit: better access:forward=yes+access:backward=no…)

118

Yes, I pointed this out in the poll above, too. But it will be close to impossible without clear instructions in the wiki, and what else can we do?

Yes just like for all the other tags in OSM. Of course someone needs to enter the correct data, but there also needs to be useful documentation about what data can and should be entered.

119

I suspect that the answer is the same as for almost everything else in OSM, go out and Map It Properly**. To help with that, you can create an overpass query that looks for a subset of the problem data.locally (that’s just one example; there are lots of others). In some cases you will be able to fix stuff while out and about, and then rerun the query to make sure you have not missed anything.

The wiki is a bit of a red herring as people can and do write what they like there, and most mappers do not refer to it while mapping (they use editor presets etc.).

** a very old meme in OSM

120

It is not, which is why I’m arguing against this tagging.

I don’t understand what you mean. Why would oneway:foot=yes disallow riding in the opposite direction :thinking: