Add bollard type to bicycle_parking quest

[Phyks]

Hi,

First of all, this might be related to #330 and #1038 where some arguments against having "bollard" type were mentionned.

Use case

There is no simple way to tag a "bollard"-type bicycle parking in StreetComplete at the moment. These are to be considered as stands. See https://wiki.openstreetmap.org/wiki/File:Bike_path_on_College_in_Toronto.jpeg for an example of bollard.

Proposed Solution

I suggest "bollard" type could be added to the bicycle parking type quest. My main arguments for it are:

• "bollard" value is relatively spread https://taginfo.openstreetmap.org/keys/bicycle_parking#values (given the fact that this type of stands is more rare than regular stands for instance).
• As a bike user, bollard offer a quite different experience from what could offer other stands. Typically, a bollard is always not very large, meaning there might be conflicts if you park one bike on each side (contrary to wide stands https://wiki.openstreetmap.org/wiki/File:Wide_stands_1.jpeg).
• Some cities are starting to turn bollard on the side of the sidewalk (for instance) into bike parking by fixing extra legs on them. These could be poor quality bike parking as the bollard was already there and not designed with bikes in mind.

Thanks!

P.S.: The same could actually be discussed for wide stands as well, which ensures your bike will lean on an horizontal metal part and which should be larger so that two bikes can be on each side without conflicts.

[westnordost]

You linked the right issues. In #1038 I explain why there is nothing beyond wheelbenders and stands.
Arguments for adding bollard would be that

• it is visually very different and users might be confused what to choose. But IMO I solved this in Add "rack" to type of bicycle parking quest #1038
• it is in effect something yet different from a stand. I am not convinced that this is the case. I do not understand what you mean with that there may be a conflict when two bikes are attached. Do you mean that oftentimes it is not possible to use a bollard for two bikes? How so? (Pictures?) And even if, there is a quest that asks about the bike parking capacity. Then this capacity will simply be lower, no?

[Phyks]

it is visually very different and users might be confused what to choose. But IMO I solved this in #1038

I totally agree the explanation is enough in the app at the moment and user would fill "stands" for such bike parkings without any problem. My worry was rather than having only "stands" available would over-simplify the diversity of bike stands out there and kind of limit the full use of "bollard" and "wide stand" type as detailed in the wiki.

it is in effect something yet different from a stand. I am not convinced that this is the case

I'm more reacting as a bike user here, which is really willing to see the diversity (from the worst to the best) of types of bike stands appear in OSM (as already documented in the wiki and used, seems like wide stand and bollard are used a few percent of the time according to taginfo). Typically, my (small) city only uses bollards but most users tagged them as stands through Street Complete. That's about 100 such potential bollards.

Simply stated, this is a bike stand (the most basic one). A bollard might be considered a bike stand and offers similar capabilities. But this is also a bike stand (might not be clear but it's very narrow, about 50cm large), yet offering a very different quality of experience. So a "stand" listed in OSM does not give much insights on the quality of locking and convenience provided.

Concerning bollards, they are usually not offering the best experience:

• Their rounded shape could make it difficult to lock both the front rear and rear wheel + frame (especially if you have a cargo or longtail bike). This is particularly clear on this picture for instance.
• There could be conflicts between two neighbouring bikes due to the fact that they are not particularly large (similar to a regular stand actually). What I mean is that there is indeed a capacity of two bikes per stand which is achievable, but this might not be the most comfortable solution. Basically, you often have to put bikes on opposite directions on each side of the bollard so that the handlebars don't conflicts. Capacity of two bikes per stand is an ideal capacity which is achievable so it should remain the capacity entered. But if you have long handlebars, a large basket behind your bike or such things, it might be less comfortable to use. Note this is not affecting stands the same way because they are usually larger so you can move the bikes back and forth compared to a bollard.
• Some cities are putting extra legs to pre-existing bollards and posts to convert them to bike parking, typically with this kind of system http://www.falco.co.uk/products/cycle-parking/cycle-stands/cyclepark/. It seems this is usually simply fixed with bolts and not soldered to the existing post, meaning the security might be lower than the one offered by a regular stand (not to mention the comfort might not be top level as it was not designed from the ground up to be a bike parking).

Concerning wide stands, this is especially useful in a city such as Paris which had regular stands in the beginning and is progressively moving towards "wide stands". Wide stands as tagged in OSM offers you a mean to be sure you can lock both front wheel and rear wheel + frame (they are long enough, this should be fine with longtail bikes as well) and they are large enough to limit the conflicts between handlebars and peripherals.

[matkoniecz]

meaning the security might be lower than the one offered by a regular stand

Note that it is not fully distinguishing them from stands. There are also some poorly secured parkings among them.

[westnordost]

Hm, you said it yourself: Bollards might negatively affect the capacity, but other stands as well, if they are too close to each other. Also, there are stands that are as narrow as bollards (the 50cm picture). So, I do not see a property here that generally distinguishes them from any other kind of stands.

In my opinion, the tagging scheme as it is documented in the wiki is a bit broken, because any kind of shape, even "bicycle trees" made it into the list there.

• some values are ambiguous (racks vs stands, anchors vs wall_loops for example)
• some values are defined by appearance, some by quality/security
• there is not any categorization, it's almost a free-form value

This is why I reduced what StreetComplete users can tag directly to two things: unsafe and safe - front wheel only and frame.

Further subcategories of these in my opinion belong into subtags, if there was a scheme for that (a wide stand, a curly stand, a wavy stand, a narrow stand and bollards are a type of stand).
But question is if they should be tagged at all. What is the value or use case to define exactly what type of stand there is? I can only think of something halfway far-fetched. The most important thing is to map whether they are reasonably usable (safe) or not (unsafe), and that is done in StreetComplete's bicycle parking type quest. Adding any more details I consider a separate task whose effort to add and maintain this data vs gain is unclear to me.

[Phyks]

So, I do not see a property here that generally distinguishes them from any other kind of stands.

What I was saying was rather that "stands" is a very generic (too generic actually) description. As it is well explained in SC, a "stand" is just something you can lean your frame on.

Further subcategories of these in my opinion belong into subtags, if there was a scheme for that (a wide stand, a curly stand, a wavy stand, a narrow stand and bollards are a type of stand).

I tend to agree with this. However, I'm not really willing to try pushing a new tagging system for "bicycle_parking" based on security vs appearance at the moment :) I was trying to get the most from the existing tagging scheme with SC :)

My main argument here was that:

• a "stand" is too generic to give a real insight on the quality of the bike parking.
• a "wide stand" might be visually oriented, but it narrows down the quality of the "stand" as well. A "wide stand" will guarantee you you can use the bike parking at full capacity (they are large enough) and you can secure both front and rear wheel + frame.
• A "bollard" (even though this is visually oriented as well) on the contrary can be considered as a "not very good stand" as usually they are quite narrow and you cannot lock both frame, rear wheel and front wheel. There might be negative effects on capacity as well.

Of course, "stand" category will be very heterogeneous as not everything would look like a "wide stand" or a "bollard". At least, for everything that looks like a "wide stand" or a "bollard", it will be tagged as such and convey more infos for the end users.

Additionnally, in my area (Paris region, France), most pre-existing stands are poor quality stands (regular stands, bollards). New bike parking added are "wide stand", so making a clear distinction (still visual, I agree that's not perfect and should be based on quality and safeness rather than look) here would help keep track of which are the newest (and then the best) spots.

[westnordost]

Okay, so the definition of "wide stand" is, that both front and back wheel can be secured(?)

However, bollard is a definition by appearance, so it does not seem to fit in there, it is not consistent.
This bike stand has similar properties as this bollard, they are both quite narrow.

The problem with the missing proper categorization in the values currently documented on the wiki page for bicycle parking types is this:
If stand is implicitly a category of "any kind of stand as long as you can lock your frame to it" but at the same time on the same level as its implicit children like wide_stand, it becomes impossible later to tell apart if one bicycle parking is just "a typical medium-sized stand" or "wasn't tagged more precisely because the user didn't care".

We have this problem at a number of different places, such as buildings (building=residential means that it can't be defined more precisely or the previous user didn't care to?) or surfaces (surface=paved means that it can't be defined more precisely because it is more or less a patchwork or the previous user thought that paved is enough "for now"?).
I want to avoid this with bicycle parking types, or at least not contribute to the problem, by clearly separating the category from the detailed (appearance-based) tags.

[brunnerpaul]

I'm commenting mainly because of the "feedback required tag".

Relevant note: https://www.openstreetmap.org/note/3519463
Image: https://westnordost.de/p/128498.jpg

I agree with two quotes from @Phyks:

Typically, my (small) city only uses bollards but most users tagged them as stands through Street Complete.

A "bollard" [...] can be considered as a "not very good stand"

Not adding "bollard" was IMO a good decision as it's basically a sub-type of stands, so it would only take up more precious screen real estate. Adding "two-tier bike rack" on the other hand was important because that's an entirely different type regarding functionality and how it looks.

My distinction between "bollard" and "stands" is based on how far the two poles are apart (most "bollards" I see have two poles). I've also tagged "stands" in cases where I would now prefer to use "bollard".

Previously, I was wondering if the distinction between the two is worth a) giving up the ease of StreetComplete on my side for that task or b) for StreetComplete to offer "bollard" additionally, only for other users to realize "wait, these were a different type?" :)

My favorite answer to that would be a third option: to have more types offered in StreetComplete Expert Edition (I see there are already additional types for some other quests there).

That way, StreetComplete would achieve this (quote from @westnordost)

I want to avoid this with bicycle parking types, or at least not contribute to the problem, by clearly separating the category from the detailed (appearance-based) tags.

... while StreetComplete Expert Edition would give users more choices without confusing the general StreetComplete user base.

@Helium314, what's your take on this? Should I open an issue for that in your repo?

[Helium314]

My first impression is that bollard is not really worth a separate answer, but then I've never actually put my bicycle to one, and it's 5th place for bicycle_parking in taginfo.
It definitely could be included in EE.

[curbia]

I would like to see bollard included, at the very least in EE. It's the dominant form of parking here. rack would also be a necessary addition for me.

As it stands, I will have to edit every single bike parking node manually at a computer because of the lack of inclusion of these tags in StreetComplete. I think SC should be aiming to not add incorrect information to the map whenever possible. I think the average user can tell the difference between these types of parking based on photos.

[mnalis]

I would like to see bollard included, at the very least in EE. It's the dominant form of parking here. rack would also be a necessary addition for me.

Can you post here a few typical pictures of each in EE @curbia (preferably with bicycle nearby) to be sure what type exactly are you talking about?

Looking at the wiki I do not see huge issue with missing bollard subtype: it seems to me that wide_stands, bollard and (some meanings of) wave are all just subtypes of stands (see both wiki description and a StreetComplete text "supports bicycle frame" which is very important part of the answer).

So using stands for bollard is IMHO not incorrect, just less precise. (e.g. like some people might tag building=house instead of building=detached - it's not incorrect, just less precise; but in any way it is improvement over building=yes).

I think SC should be aiming to not add incorrect information to the map whenever possible.

Absolutely. If user is not sure, they should say Other answers / Can't say; and leave a note (preferably with a picture) -- and not selecting an random incorrect answer. Hopefully most of users would behave in that way, but there is no guarantee.

I think the average user can tell the difference between these types of parking based on photos.

Sure; but I still don't still what is fundamental difference from stands? There are many types of stands, and the one of more popular ones shown in the picture (called "staple rack" in North America) is just one of them. What they all have in common is that they support whole bicycle frame, and allow locking both frame and wheel to it. If you can do that with bicycle_parking=bollard, then it is a subtype of bicycle_parking=stand. E.g. all parkings in picture below are bicycle_parking=stand, even if they are not staple-shaped:

Would bollard also fit in that picture; or is it fundamentally different (and how?)

I think main problem with adding subtypes of stands is that if you add one such exception, you really should add all of them; otherwise one risks creating more confusion than solving problems. And adding all bicycle parking type is impractical and creates stress for mappers who are then unable to decide which one to use.

rack would also be a necessary addition for me.

Looking at the wiki, the bicycle_parking=rack tag seems ill-defined: it seems to be used both for wheelbenders and coathangers, as well as for stands sometimes (see North American usage of "staple rack" to mean bicycle_parking=stand instead). I would avoid using that value and use some other more precisely defined tag.