What does "Usable by" mean?

General talk Tagging general discussion
1

The previous discussions here and here (a.o.) on smoothness tagging practices made it clear that there are two practices that differ in the way they interpret the “Usable by” column in the Values table. I think it’s time to start a poll to finally make an end to the divergent practices, so that a certain smoothness value means the same worldwide and doesn’t depend on the mapper.

One interpretation focusses on the better smoothness values (down to very_bad) and the vehicles described in the Usable by column. Thin_rollers are mentioned in the excellent column. However, such vehicles could also use the surfaces in the good column, although with difficulty (slow down, be more attentive) and reduced comfort (thanks @Matija_Nalis for pointing this out). Similarly, the thin_wheels vehicles in the good column could also use intermediate surfaces though with difficulty. This approach was used by the Berlin Verkehrswende group and then by Streetcomplete when the smoothness quest was developed. @Westnordost then created the Smoothness Gallery as a result of the hard work of the StreetComplete developers. I would like to call this the “Comfortably usable” practice.

The other interpretation focusses on the phrase “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)” that was added to the wiki soon after the Key was approved in Nov 2008. It focusses on the worse smoothness values (from bad and worse), where each vehicle mentioned in the “Usable by” column is required to be able to pass the way (though with care and discomfort), and a surface of the next worse smoothness column is not passable (not even with great difficulty and discomfort) by that vehicle. I would like to call this the “Just passable” practice.

The two practices diverge especially where they overlap, i.e. in the range from intermediate to very_bad.

Which practice do you think should be used to evaluate smoothness?

  • Comfortably usable
  • Just passable
0 voters
2 Likes
2

My arguments for voting “Just passable”:

  • From the beginning of the establishment of the smoothness key, it was intended that “usable” should be interpreted as “without significant risk of damage or injury”. Comfort was never mentioned. “Just passable” is the original intent of smoothness

  • “Passable” is a much more objective and verifiable criterion than “comfortable”, which is very personal. Anyone who drives a normal car with ground clearance >14 cm on a road and damages the undercarriage can objectively and verifiably conclude that smoothness=very_bad or worse. Any mapper who sees that a certain bad road is used by a few normal cars can objectively conclude that smoothness=bad or better, while he can’t observe how many normal cars avoid that road because it is uncomfortable to drive on. A mapper who sees someone riding a racing bike fall or damage the wheel rim can conclude that smoothness=intermediate or worse. It is much more difficult to objectively determine for the average mapper how many racing bikers avoid a way because it is uncomfortable. It is very personal what people find “comfortable”: some find it very important; others don’t mind roughing it a bit, and there are masochists…

  • All ways with smoothness=bad and worse are uncomfortable, and the vehicle used doesn’t matter much. Consider this picture, which is in the very_bad (Car with high clearance) column in the present Gallery. However, it is passable by a normal car (though slowly), and using a car with high clearance doesn’t make it more comfortable (you would have to drive equally slowly).

    1200px-Kamieniec_Szalejow_Gorny_road
    1200px-Kamieniec_Szalejow_Gorny_road1200×900 283 KB

  • We should not be the ones to determine what level of comfort is acceptable, but leave that to the data end user. A routing app could have a setting by which an end user could choose themselves which comfort level is acceptable to them. A user who doesn’t like to cycle on smoothness=intermediate ways could then set the app to allow max. 100 m on such road (or max. 1% of the total route, or max. 30% longer route to avoid smoothness=intermediate ways, etc.). A user who doesn’t mind cycling on such ways could set the app so no detours to avoid them are allowed.

  • The practice is already in use for the “bad end” of smoothness values, so why should there be a different practice for the “good end”? We could edit the wiki so that excellent is for all_wheels, good is for thin_rollers and intermediate is for thin_wheels. This is a relatively small edit to unify the different tagging practices.

Now off to do some mapping :wink:

3

My choice of answer was my intuitive gut decision as I have always used and understood this tag.

After that, I looked very, very closely at the OSM wiki, especially the very long discussion on the smoothness Talk page and the even longer discussion on the 2014 mailing list.
Initially, the aim was probably to show whether a road is “comfortable” or rather “uncomfortable” to “dangerous” for a group of means of transportation. The point was that the map shows ways in which one must maintain “caution”. So far so good, but still very vague. The alternative proposals with “trafficability” and “usability” has also been suggested. See also: User talk:Davo - OpenStreetMap Wiki

I would say that the consensus of the whole discussion in the wiki is just about how “smooth” a road is / “smooth” enough to drive on. It should be “good usable for traffic” - which I would say is more than “just usable”. It is also often said that a vehicle should be “comfortable”, indicating the level of comfort and suitability for different types of vehicle, rather than simply indicating whether a road is just passable. Examples were given that a way tagged as excellent implies a high level of comfort for thin-wheeled vehicles, while a way tagged as bad is still passable by normal cars, but with reduced comfort.

After reading this, my understanding is that the consensus within the OSM community leans toward interpreting the tag as indicating how “comfortably usable” a way is for different types of vehicles, rather than simply whether it is “just passable”.

Therefore, I will keep my vote unchanged. Personally, I can put myself in the shoes of a racing cyclist the best, and when I look at the standard picture from the wiki for intermediate, then I can ride on this road, so it is definitely “just usable” for me on my road bike. But it is certainly not comfortable, and in the end, these roads are not the reason why I get on my road bike at all.

Even with my car driver’s glasses on and looking at the current picture for “bad”, I see it the same way. Yes, I can drive there with my car, but if I only had to expect roads like that, I would leave my car at home.

I agree with you that in the original proposal, it just says “usable by”. The question then is, what was the intention of the word “usable”? In the end, it just means “you can use it”, without any reference to how you can use it. Should the tag actually mean “How can you use it? Good, bad, horrible…”?

The most basic definition of “usable” could be passability. In other words, a vehicle can simply traverse the path without immediately suffering a mechanical failure or becoming entirely stuck. The rider should be able to maintain forward motion without excessive difficulty.
But I would say for example road bikes are designed for smooth, paved surfaces. A surface classified as “usable” should ideally allow the rider to pedal efficiently without excessive discomfort. The presence of potholes, cracks, loose gravel, or rough surfaces would make the ride unpleasant or even dangerous. A distinction should therefore be made between whether something is ideal, acceptable, barely usable or even unusable.
I therefore interpret “usable” as something where you can drive efficiently and safely.

My conclusion / my personal, most logical interpretation of “usable by” is a surface that allows continuous travel with reasonable comfort and without excessive risk.

5 Likes
4

My argument against voting “just passable”: I’ve taken my Honda Civic on a number of high-clearance roads. If you go slow enough (walking pace or slower), the suspension doesn’t bounce and impale your undercarriage on pointy rocks, and a narrow, agile car can go around many of the obstacles that nominally require higher clearance.

I would not recommend these roads to the average driver because of the hazards. For example, when driving back from one hike, I passed a similarly low-clearance vehicle that had left part of its oil pan on a rock the driver hadn’t spotted in time.

1 Like
5

My take on this: just passable is no more objective than comfortably passable. Driver skill and urge take the lead there, but this does not make it any more reproducible.

4 Likes
6

I use smoothness as a guideline for general public use, rather than as a measure of a vehicle’s or rider’s limits. For motorcycle users, whether a route is barely or comfortably passable largely depends on their vehicle, skill, and experience. Since most users will likely fall into the lower skill category, “Comfortably passable” seems like a more appropriate fit.

6 Likes
7

I really don’t want to be misunderstood as a finger-pointing gesture, and I am really interested in your opinion. May I politely ask you, @Mateusz_Konieczny, to share your perspective about “just passable”? Thank you!

1 Like
8

Does that mean you (and others) answered that question as if it was asking “what is your own current tagging practice?” I meant to ask what the tagging practice should be. Depending on the results of this poll, some will have to change their tagging practice, which is of course uncomfortable :wink: I hope that didn’t play a role.

Thanks for your effort; I didn’t go that far back.

This apparently means that it was found acceptable that the criterion changes from “Comfortably usable” for the good end of smoothness to “Just passable” for the bad end.

Is this because you cycle for pleasure? That’s very personal. Would it be different if you had to take that road to get to work? Would you buy a more sturdy bicycle for it then?

That would be an indication that it’s smoothness=bad

That would be an indication that it’s smoothness=very_bad. Assuming this is about the same road, it shows that “Just passable” is not a fully objective criterion either. It could be that you are an exceptionally good driver, or that the driver of the other car is exceptionally bad. You’d have to compare the number of broken down “normal” cars with the number of drivers of “normal” cars who do take the risk to drive that road. However I think “Just passable” is more objective than “Comfortably usable”.

That’s what you should have in mind, it’s what the “Will most people be happy…" phrase in the wiki is trying to say.

I think avoidable obstacles should not be taken into account, but of course it takes some skills to do it. That smoothness depends on the width/agility of the vehicle is a major weakness of the smoothness key.

I do think that driver skill varies less than acceptable comfort level. Both could be taken into account by a user of a routing app when selecting what is acceptable to them. In my arguments post (no. 2), where I write “A user who doesn’t like to cycle…” could be replaced by “A user who doesn’t have the skills to cycle…”

9

Both would be the same answer for me. :slight_smile:

Actually both, in my current (ideal) world, I have pleasure in my commute :slight_smile:

If all the roads in my area were in an incredibly bad condition, then I would probably swap my racing bike for a fat bike or a model similar to a fully. :mountain_biking_man:

1 Like
10

It’s what it should be: just passable would simply be wrong.
And if I’m in doubt, I don’t use the better option. I try to think of the average cyclist, not one of the 5% who are technically great and not one of the 5% that maybe are technically even less fit than me.

I’m commuting, I’m cycling for pleasure too. And I’m a dedicated user of wandrer.earth what gives me all sort of ways, all options of smoothness from excellent to impassable, all options of trail_visibility too :smiley:

11

My 2-cents for what it’s worth. I think smoothness should remain objective as sometimes a subjective ‘gist’ is still better than nothing (taken with a grain of salt).

If you want true non-subjective this there is International Roughness Index (IRI) which is quantitative. Yes, I know it’s not something that OSM has tags for but cities are paying for this but it’s mostly paved road quality and repair/replacement planning to save money.
Then for vehicles rather than listing vehicles by smoothness list suggested tags to include to address smoothness we have plenty of options…

If we want vehicles specific rating don’t we have those in part?
Vehicles: tracktype=grade1-5 + 4wd_only=yes/no
Mountain Bikes: mtb:scale
Hiking: sac_scale
Horseback: horse_scale
Motorcycle: motorcycle:scale
Dirtbiking: dirtbike:scale

12

I read somewhere, 90% of drivers consider themselves to belong into the top 10% tier of driving skill. Unfortunately too tired to look up the source of that quote. Did I dream that up?

13

No, I’ve also seen similar sorts of figures quoted.

1 Like
14

Hard to believe, especially when dealing with rough/unpaved sections.

15

Questionable. At least the study I know of is about the correlation between one’s own assessment and an expert assessment. The study identified that 94.8% of the participating drivers rated their own performance to be better than the assessments of the expert rating.
Source

Last year, the automobile club in Germany commissioned a representative online survey in which 82% stated that they were good to very good drivers. (29% of them said very good).
Whether “good to very good” should be defined as the top 10% remains questionable, but the ADAC did not do so.

1 Like
16

IMHO, the purpose of the smoothness tag is to allow a good decision about which route to take with a given vehicle. Based on that, I think the correct answer here would be without significant risk of damage or injury and without requiring any special skills.

With this definition, the decision for any routing application is clear: if the smoothness is below the minimum required for the selected vehicle, don’t go there. For all other values of smoothness, the router can (and probably does) still apply different penalty values, so if it could lead your car across a bumpy (smoothness=bad) shortcut, it will do so to save you 15 minutes, but maybe not if it makes you just 1 minute faster.

The pure “Comfortably usable” is ambiguous for routing algorithms: If you can go comfortably with your car on smoothness=bad, should it consider smoothness=very_bad as an option when it’s much shorter?

1 Like
17

is highly dependent on the driving skills of the individual and therefore not suitable for the definition of the smoothness classes

For comfortable, however, I would expect a smaller spread.

18

I guess two words “just passable” is not enough to transport
“without significant risk of damage or injury”

With this definition in mind I may had decided differently.

19

What’s the argument against adding both “comfortable for” and “just about usable for” to the table?

20

looking at this again I have no idea what I meant here

I changed my vote

1 Like