Smoothness on highway=steps

General talk Tagging general discussion
, ,
1

As for the footnote, I’m actually just asking to learn if this is a known and relatively widespread feature. Given the scarcity of contributors to conduct detailed large-scale surveys in Brazil, we should direct these available resources toward mapping activities that generate more practically useful accessibility data more quickly.

I agree. In my comments, I assumed mapping smoothness on highway=footway rather than highway=steps. [Off-topic] If smoothness on highway=steps is relatively uncontroversial, could we document, for the sake of clarity, that it refers to the evenness of the surface within each step, rather than along all of them?

Inaccessible sidewalks in Brazil
2

I disagree. smoothness describes the roughness of a way along its whole length for wheeled vehicles, so for highway=steps, smoothness would be proportional to the step height. I do agree that it’s a bit odd to tag smoothness on steps, but it could make sense on the steps-like sidewalks that this post is about.

2 Likes
3

It surely looks like so. The intro paragraph in the wiki has been there for years uncontested and says:

smoothness provides a classification scheme regarding the physical usability of a way for wheeled vehicles, particularly regarding surface regularity/flatness. Usability by a particular type of vehicle means that the vehicle can traverse the route without significant risk of damage (e.g., to vehicle undercarriage/drivetrain) or injury (e.g., by falling). The tag can be applied to all drivable ways and areas: highway=*, parking areas, beaches, etc.”

However, searching for “smoothing steps” does not lead me to a specific discussion about this combination of tags. Maybe someone could argue that highway=steps isn’t really drivable most of the time by most people on most vehicles.

4

Well. Opinion of how things “should properly be mapped in theory” is not required to match how things “are actually being mapped in practice”; and I am of somewhat strong belief that the latter is MUCH more important in practice (i.e. if you are data consumer or such).

See this post for my take on different “Factions” of belief regarding OSM tagging. – I’ve put myself mostly in descriptivist category[1].

And if we do look how smoothness=* is actually tagged on highway=steps in practice (taginfo) we find out that (out of 10.7k) of them, about:

  • 3.9k are smoothness=good
  • 2.4k are smoothness=intermediate
  • 1.9k are smoothness=excellent
  • 1.4k are smoothness=bad
  • the rest are below taginfo detection threshold

If the smoothness on highway=steps were mapped according to your beliefs, I would expect vast majority of highway=steps to have smoothness=* of levels impassable and very_horrible and maybe some small amount of horrible ones)

But they are not, and instead ~90% of the steps are tagged with type of smoothness where it would be quite reasonable (if not preferred!) to drive a car on should your interpretation hold (certainly excellent and good and intermediate could practically never match your definition, and even bad is defined as being just fine to drive on using “a normal passenger car with >14 cm clearance”).

While I am not expert in cars by any means, my opinion (and practical experience in my little corner of Europe) is that most (if not all) highway=steps are usually not suitable to driving a regular passenger cars on (at least I have never seen one attempt to do so in real life), much less driving them on steps with the same comfort and speeds as on say motorway (as smoothness=excellent/good would suggest in “your” interpretation).

Note that highway=steps documents step:height=* (and step:length=*) should one feel inclined to map that information, so using smothness=* for that purpose would be redundant.

TL;DR: So, it seems to me that documenting actual usage (and disagreements of some members of community with current tagging) of smoothness=* on highway=steps in wiki is in order, as wiki documentation is clearly not matching the reality. Would you agree?

  1. and I’d probably put you @rhhs in prescriptivist category – would you agree? ↩︎

5 Likes
5

I’d call myself a pragmatist: I think our ultimate goal should be to make a map that’s increasingly useful to end users, and for that we need to innovate our tagging schemes. For this to happen, we need a combination of anarchy/chaos in the beginning (this promotes creativity) and organisation later on. It’s like brainstorming: you start by being very free to speak out your ideas, no matter how stupid they may seem, but later on you have to look at the ideas that were suggested and organise them, filter out the good ones, decide if they’re worth implementing and make an action plan to implement them.

I think it’s a good thing that a couple of years ago some people got the idea to create the smoothness tag and started using it (and ignored the verifiability complaints) to find out in practice how it works best. But then new user rhhs came along and started documenting it in detail in the wiki, ignoring the fact that this is usually first discussed on this forum (he didn’t know about it). I think it is now a more useful tag than it was: it’s good to have it to get a rough indication of how usable a way is, but it’s even better if a map end user can rely on it that it means the same in Zagreb or Sofia.

It would be good to create a separate thread to discuss this.

I’m surprised there are so many, and that the majority of them are with quite good smoothness values. I’m wondering how many of them were created when a section of steps were split off from a path/footway with a smoothness tag, and the mapper forgot to remove the smoothness tag from the steps he created… It would be good to find out how many of them are connected to a path/footway with the same smoothness tag value…

(would be good to split smoothness on steps to a separate thread too)

2 Likes
6

This way Way: ‪Kerkstraattrap‬ (‪903156497‬) | OpenStreetMap can now be tagged smoothness=very_bad (or worse): Verstrooide Nederlander rijdt per ongeluk van statige Kerkstraattrap in Blankenberge | VRT NWS Nieuws :joy:

afbeelding
afbeelding600×822 120 KB

2 Likes
7

In fairness, as a Dutch person, the driver might not be used to such differences in elevation? ;)

2 Likes
8

You keep using that word - I don’t think it means what you think it means.

According to wikipedia:

Pragmatism is a philosophical tradition that views language and thought as tools for prediction, problem solving, and action, rather than describing, representing, or mirroring reality

So, are you really trying to say that you intentionally do NOT want the OSM wiki to be “describing, representing or mirroring reality (of how we tag in OSM)” ??

If so, I think more understandable lingo in today’s word would be to say that one “intends to weaponize wording on wiki to accomplish their own agenda (i.e. promote certain ways of tagging they prefer)”?

That would definitely fall under my definition of “prescriptionist”, yeah.

Sure, I like that. We have a thing just for that in OSM; and it is called proposal process. But you should use it on new tags, and not on ones which already have different meanings, for which such changing of meaning will inevitably lead to irrevocably skunking them (which is generally considered quite negative outcome).

They still can’t. Wiki is not some kind of magical make-a-wish-come-true machine, in which you put words and they become reality. Not even if you keep wishing really hard. But I’ve tried to explain that many times, so if it hasn’t come through yet, I have doubts that repeating it will make it so.

Re: smoothness saga...

As you might have noticed, even I[1] have given up on even trying to participate in 37th reincarnation[2] of “how to redefine smoothness tag meaning again” threads. If the intention was to “win by wearing out anybody with opposing views trying to participate”, it certainly succeeded, so I suppose congratulations are in order.

It is not my intention to make you sad, but if we put mappers into few rough categories:

  • those that only glance at a picture provided by their editor (or, more likely, used to do it in the past before it become mechanical memory)
  • those who bother to read short (8 words or less) description provided by their editor (or, more likely, used to do it in the past before it become mechanical memory)
  • those who have memorized what some tag does once upon a time (like, 10 years ago - maybe in some editor help, or forum, or even longish one on the wiki), and don’t have any intention of ever looking it up again (much less re-learning new definitions)
  • people who subscribe to wiki changes and/or obsessively re-read their pet tag definition every few weeks, and re-learn by heart how much clearance in cm each car must have for certain smoothness in this week newest edition

… then I’d said that the last category is at most 1% of mappers (and that is likely overgenerous).

And even that last category, that does follow and update their new tagging to newest wiki claims – do you think they ever go back and re-tag their last few decades of using that tag when its meaning is changed on the wiki?

Even if they wanted to (and were willing to spend many months of their life retagging), do you really think they can remember what was the exact clearance in cm for that track they passed once 17 years ago? It’s inconceivable that even a tiny part of that minority would be able to do that.

And thus, you’d inevitably as a result have some instances of tag which mean old thing, and some which mean a new things. And no way to know which are which. Congrats, your “helpful” changes to the wiki have managed to skunk that tag. Brave New World, indeed.

To address your example: when you carefully measured the holes to be at most 19cm thus passing a clearing of certain model of a car, and so you marked it as smoothness=horrible in Sofia; that is not at all the reasoning I had when I mapped some other path as smoothness=horrible in Zagreb. It was more of “this would really suck for my trekking bike” here. And no, adding even more changes to the wiki would not change that.

And, no, the idea of rereading the wiki before clicking that answer in StreetComplete was not even on the farthest backburner in my mind. And I really doubt I’m in ignorable minority with such behaviour; to the contrary, I’d guess it is by far most popular behaviour in reality.

But I can’t prove those numbers. We could put a vote or something to see what at least one biased part of mappers (i.e. Discourse crowd) thinks, but I really don’t think that even that would change your mind. Would it?

TL;DR: changing meaning of popular existing tags is impossibly hard. The best outcome one can hope is that the tag becomes only somewhat skunked, and not totally ruined.

Please, if your intention is moving OSM forward with better definitions, let’s do it with new tags only (and eventually deprecating old tags when not needed anymore), and not changing definitions of existing popular tags contrary to how those tags are used in reality currently. Otherwise, it is not helping, but is IMHO borderline vandalism.

  1. who am likely among top 0.01% mappers having patience for such things ↩︎

  2. or 137th, I have really lost count ↩︎

3 Likes
9

According to wiktionary a pragmatist is 1. One who acts in a practical or straightforward manner; one who is pragmatic; one who values practicality or pragmatism. I think it is obvious from the rest of the paragraph that this is the meaning I intended so I wonder why you picked meaning no. 3. Because it was the first result showing up in your search engine and suited your intention to make me ridiculous, maybe?

The rest of your epistle is a repetition of your false assumptions about my intentions that I already explained in a few earlier posts and won’t bother to repeat here. If you want to discuss them again, please start a new thread or send me a PM because it is off topic here. Please assume good faith and limit your use of bold and italic text and irrelevant links to Youtube so we can have a good discussion in mutual respect.

10

Talk about assuming good faith :frowning:

If my intention really was that, wouldn’t I have rather quoted 4th (obsolete) definition of pragmatic? i.e.:

Interfering in the affairs of others; officious; meddlesome.

But I didn’t and I don’t, because belittling you is not (nor it was at any point) my intention.

If anything, my intention is to try to channel your energy and enthusiasm about the project in a way that would be aligned with how other are choosing to use it, instead of a way you personally envision, but majority of others dislike (and map differently), and thus your misguided efforts resulting in avoidable damage the project while you continue to believe is helping it.

Actually, the rest of the paragraph was what convinced me otherwise.

SO, to try to clarify that I didn’t have dishonest or rude intentions, but simply picked only meaning that matched (while asking you explicitly to confirm if you that is what you meant):

Half of the meaning of (1) (“one who values […] pragmatism”), and most other meanings (2) “One who acts in response to particular situations rather than upon abstract ideals[1]”, (3) “One who belongs to the philosophic school of pragmatism” and (4) “An advocate of pragmatism” all relate to that same wikipedia article I quoted.

In my opinion, surely you couldn’t have be thinking of one half of first definition “One who acts in a practical or straightforward manner” when what you’re proposing in the same post is the complete opposite of that (and yet matches other 3.5 definitions - somewhat similar to “prescriptionist”[2] very well)?

To wit:

  • “Practical manner” would be to map smoothness on steps as everyone else[3] seems to be doing - and not on getting involved in convoluted and prolonged personal crusade to have “legislature” passed that would support your world view how you would prefer them to be mapped, even if practical result of such action would unavoidably would’ve been catastrophic (as detailed in rest of my post which you happily decided to summarily ignore")
  • “Practical manner” would also involve just quickly eyeballing the surface compared to picture on phone, instead of carrying a measuring device with yourself and getting out of the car to precisely the measure each hole, so you can match them correctly to your arbitrarily-made table.

Which part of your trying to prescribe on wiki how others should use certain tag (if you accept you are doing that?), even after it is explained that it is extremely unlikely to influence anyone but very small minority of people, would you describe as “acting in practical or straightforward manner”?[4]

Thus, given the obvious ambiguousness of that word “pragmatist” (and related words), might I suggest that we skip from using it completely, and stick to terms “proscriptionist” and “descriptivist” to describe whether one believes the wiki should be proactive or reactive respectively[5] ?

I am acting in good faith, and am in fact very convinced that you’re trying to do the right thing and improve OSM ecosystem; the point I am trying to make across is that the specific way[6] you sometimes seem to choos in order to accomplish that (i.e. making wiki prescriptionist instead of matching reality of tagging) is counterproductive.

If I did not think you were intending to do good, we wouldn’t be discussing at all[7]. But I as I say, I strongly believe you sincerely want to help.

IMHO, You just need little more effort to actually listen to what others are saying as the whole thing; instead of using it as a prelude to composing the answer explaining why you don’t agree with them and intend to ignore them.

Why in the world would you ask for that? They are intended to emphasize certain parts which are most important. Removing them would just reduce legibility of the message significantly[8]

Ok. I’d tried to add those funny videos to reduce tension, show good faith and attempt to lighten the atmosphere, but if you dislike that, I can keep to just hard facts. (Though I’d note that that funny picture of the traffic accident you posted was pretty much in the same vein, and I’m not complaining - to the contrary…)

Wait; please tell me you’re not really trying to skip addressing the entirety of the important stuff in my post with such a blatant wave of a hand (without even attempting to show any arguments), just so you can complain about few tangential sidenotes you disagreed with?!

Ok. So I’ll start by apologizing if what I wrote felt rude or belittleling. That was not my intention, but I do think that wiki-fiddling is not something that can change how most mapper map.

I’m sorry if you found that video insulting or rude, I think it shows (in funny way. C’mon, you really didn’t find it funny? :crying_cat: ) that OSM wiki, just like coffee table in that sketch, does not (and cannot) have magical qualities that it seems you are assigning it (i.e. ability to influence how vast majority of mappers map just by writing something there that you wish it was used, instead of how it is actually used in reality).

So, would you now be so kind to respond to the actual meat of my post, instead of ignoring all that mattered[9] and concentrating on mostly irrelevant sidepoints like formatting and choice of specific dictionary definitions?

Please, if you feel you already answered some question in full in previous post, please quote relevant part of that post as your answer. That way you’ll make me be humble and apologize if I missed it (and/or ask for further clarification if I feel it only partially addresses, or completely misses, the question I was asking or claim I was making).

Or would you prefer that I repeat main questions/claims in separate post, but without those links to dictionaries and videos?

OK, I’ve asked the mods to split several posts to a new thread[10], and it seems to have been done now!

  1. pushing for “correct” tag interpretation that matches your worldview but mismatched tagging reality is definitely persuing abstract ideals IMHO ↩︎

  2. which you, in my opinion, very much seem to be in this and other threads ↩︎

  3. seem my previous posts for taginfo proofs ↩︎

  4. If that is indeed the meaning you have intended to use? ↩︎

  5. or if those two words does not suffice, we may use legal practice of inventing a new word starting with Capital letter (or better yet, in ALL-CAPITALS), and provide at least paragraph-long description of what word means to us exactly, before starting to use it in Captalized form exclusively – in order to avoid misunderstandings ↩︎

  6. i.e. treating wiki as make-wish-come-true edifice ↩︎

  7. instead it would be wikiadmins or mods or DWG or OSMFboard getting reports ↩︎

  8. just like would removing formatting, uppercase letters, interpunctions etc. ↩︎

  9. that would not be very pragmatic, if we look only at first part of that 1st definition ↩︎

  10. FYI, you can done it yourself too should the need arise in the future, using that flag icon under three-dots-menu ↩︎

11

Anybody here who’s smart enough (I’m not) to test my hypothesis on why there are so many highway=steps with good smoothness values? For instance by making a query that analyses how many of the thousands of cases that @Matija_Nalis found share a node with a highway with the same smoothness value? That would be an indicator that the steps were split off from it without removing the smoothness tag.

I find it hard to imagine that anyone would map a new highway=steps and intentionally add a smoothness tag.

12

I would do this to distinguish between cases where steps are flat or not at all

13

Unfortunately I am not either. If it is within realms of overpass (which I am not sure about), I’d guess it would likely be very expensive there. So some other tool (maybe postpass? or duckdb on highway=*-only planet extract?) would probably be better idea. :man_shrugging:

Anyway, I’ve looked at StreetComplete git history, and it seems the ambiguity of smoothness=* on highway=steps was well known even back then in 2021, as the quest explicitly skips asking smoothness=* there (list variable storing types of highway where smoothness should be asked is called ALL_PATHS_EXCEPT_STEPS, so it seems quite intentional)

(Also, in StreetComplete, if users splits the way and convert part of it to steps, SC will remove smoothness=* tag from it, so it at least wouldn’t be creating that particular issue).

14

This Overpass query can at least visualize it.

An Overpass query like the one below returns mostly cases where highway=steps with smoothness=good are not connected to any other highways (except steps) with smoothness=good.

way[highway=steps][smoothness=good]({{bbox}})->.steps_good;
node(w.steps_good)->.steps_nodes;
way(bn.steps_nodes)[highway][highway!=steps][smoothness=good]->.connected_good;
node(w.connected_good)->.highway_nodes;
way(bn.highway_nodes)[highway=steps][smoothness=good]->.matched_steps;
(
  .steps_good;
  - .matched_steps;
);

Some extra work is needed to remove highway=steps that were split into many parts. These appear to be rare though. Then the query has to be adapted and repeated for every smoothness value.

From what I’ve seen, both situations occur in the data.

2 Likes